2021

  • K. Wiktorowicz, T. Krzeszowski, and K. Przednowek, „Sparse regressions and particle swarm optimization in training high-order Takagi-Sugeno fuzzy systems,” Neural Computing and Applications, vol. 33, iss. 7, p. 2705–2717, 2021. doi:10.1007/s00521-020-05133-w
    [BibTeX] [Abstract]
    This paper proposes a method for training Takagi-Sugeno fuzzy systems using sparse regressions and particle swarm optimization. The fuzzy system is considered with Gaussian fuzzy sets in the antecedents and high-order polynomials in the consequents of the inference rules. The proposed method can be applied in two variants: without or with particle swarm optimization. In the first variant, ordinary least squares, ridge regression, or sparse regressions (forward selection, least angle regression, least absolute shrinkage and selection operator, and elastic net regression) determine the polynomials in the fuzzy system in which the fuzzy sets are known. In the second variant, we have a hybrid method in which particle swarm optimization determines the fuzzy sets, while ordinary least squares, ridge regression, or sparse regressions determine the polynomials. The first variant is simpler to implement but less accurate, the second variant is more complex, but gives better results. A new quality criterion is proposed in which the goal is to make the validation error and the model density as small as possible. Experiments showed that: (a) the use of sparse regression and/or particle swarm optimization can reduce the validation error and (b) the use of sparse regression may simplify the model by zeroing some of the coefficients.
    @article{Wiktorowicz2021,
    abstract = {This paper proposes a method for training Takagi-Sugeno fuzzy systems using sparse regressions and particle swarm optimization. The fuzzy system is considered with Gaussian fuzzy sets in the antecedents and high-order polynomials in the consequents of the inference rules. The proposed method can be applied in two variants: without or with particle swarm optimization. In the first variant, ordinary least squares, ridge regression, or sparse regressions (forward selection, least angle regression, least absolute shrinkage and selection operator, and elastic net regression) determine the polynomials in the fuzzy system in which the fuzzy sets are known. In the second variant, we have a hybrid method in which particle swarm optimization determines the fuzzy sets, while ordinary least squares, ridge regression, or sparse regressions determine the polynomials. The first variant is simpler to implement but less accurate, the second variant is more complex, but gives better results. A new quality criterion is proposed in which the goal is to make the validation error and the model density as small as possible. Experiments showed that: (a) the use of sparse regression and/or particle swarm optimization can reduce the validation error and (b) the use of sparse regression may simplify the model by zeroing some of the coefficients.},
    author = {Wiktorowicz, Krzysztof and Krzeszowski, Tomasz and Przednowek, Krzysztof},
    doi = {10.1007/s00521-020-05133-w},
    isbn = {0123456789},
    issn = {14333058},
    journal = {Neural Computing and Applications},
    keywords = {Fuzzy systems,Least squares approximation,Particle swarm optimization,Sparse regression},
    number = {7},
    pages = {2705--2717},
    publisher = {Springer London},
    title = {{Sparse regressions and particle swarm optimization in training high-order Takagi-Sugeno fuzzy systems}},
    volume = {33},
    year = {2021}
    }

2020

  • T. Krzeszowski and K. Wiktorowicz, „Combined Regularized Discriminant Analysis and Swarm Intelligence Techniques for Gait Recognition,” Sensors 2020, Vol. 20, Page 6794, vol. 20, iss. 23, p. 6794, 2020. doi:10.3390/S20236794
    [BibTeX] [Abstract] [Download PDF]
    In the gait recognition problem, most studies are devoted to developing gait descriptors rather than introducing new classification methods. This paper proposes hybrid methods that combine regularized discriminant analysis (RDA) and swarm intelligence techniques for gait recognition. The purpose of this study is to develop strategies that will achieve better gait recognition results than those achieved by classical classification methods. In our approach, particle swarm optimization (PSO), grey wolf optimization (GWO), and whale optimization algorithm (WOA) are used. These techniques tune the observation weights and hyperparameters of the RDA method to minimize the objective function. The experiments conducted on the GPJATK dataset proved the validity of the proposed concept.
    @article{Krzeszowski2020,
    abstract = {In the gait recognition problem, most studies are devoted to developing gait descriptors rather than introducing new classification methods. This paper proposes hybrid methods that combine regularized discriminant analysis (RDA) and swarm intelligence techniques for gait recognition. The purpose of this study is to develop strategies that will achieve better gait recognition results than those achieved by classical classification methods. In our approach, particle swarm optimization (PSO), grey wolf optimization (GWO), and whale optimization algorithm (WOA) are used. These techniques tune the observation weights and hyperparameters of the RDA method to minimize the objective function. The experiments conducted on the GPJATK dataset proved the validity of the proposed concept.},
    author = {Krzeszowski, Tomasz and Wiktorowicz, Krzysztof},
    doi = {10.3390/S20236794},
    journal = {Sensors 2020, Vol. 20, Page 6794},
    keywords = {biometrics,gait recognition,grey wolf optimization,particle swarm optimization,regularized discriminant analysis,whale optimization algorithm},
    month = {nov},
    number = {23},
    pages = {6794},
    publisher = {MDPI},
    title = {{Combined Regularized Discriminant Analysis and Swarm Intelligence Techniques for Gait Recognition}},
    url = {https://www.mdpi.com/905090},
    volume = {20},
    year = {2020}
    }
  • K. Wiktorowicz and T. Krzeszowski, „Approximation of two-variable functions using high-order Takagi-Sugeno fuzzy systems, sparse regressions, and metaheuristic optimization,” Soft Computing, vol. 24, iss. 20, p. 15113–15127, 2020. doi:10.1007/s00500-020-05238-3
    [BibTeX] [Abstract] [Download PDF]
    This paper proposes a new hybrid method for training high-order TakagiSugeno fuzzy systems using sparse regressions and metaheuristic optimization. The fuzzy system is considered with Gaussian fuzzy sets in the antecedents and high-order polynomials in the consequents of fuzzy rules. The fuzzy sets can be chosen manually or determined by a metaheuristic optimization method (particle swarm optimization, genetic algorithm or simulated annealing), while the polynomials are obtained using ordinary least squares, ridge regression or sparse regressions (forward selection, least angle regression, least absolute shrinkage and selection operator, and elastic net regression). A quality criterion is proposed that expresses a compromise between the prediction ability of the fuzzy model and its sparsity. The conducted experiments showed that: (a) the use of sparse regressions and/or metaheuristic optimization can reduce the validation error compared with the reference method, and (b) the use of sparse regressions may simplify the fuzzy model by zeroing some of the coefficients.
    @article{Wiktorowicz2020a,
    abstract = {This paper proposes a new hybrid method for training high-order TakagiSugeno fuzzy systems using sparse regressions and metaheuristic optimization. The fuzzy system is considered with Gaussian fuzzy sets in the antecedents and high-order polynomials in the consequents of fuzzy rules. The fuzzy sets can be chosen manually or determined by a metaheuristic optimization method (particle swarm optimization, genetic algorithm or simulated annealing), while the polynomials are obtained using ordinary least squares, ridge regression or sparse regressions (forward selection, least angle regression, least absolute shrinkage and selection operator, and elastic net regression). A quality criterion is proposed that expresses a compromise between the prediction ability of the fuzzy model and its sparsity. The conducted experiments showed that: (a) the use of sparse regressions and/or metaheuristic optimization can reduce the validation error compared with the reference method, and (b) the use of sparse regressions may simplify the fuzzy model by zeroing some of the coefficients.},
    author = {Wiktorowicz, Krzysztof and Krzeszowski, Tomasz},
    doi = {10.1007/s00500-020-05238-3},
    issn = {1433-7479},
    journal = {Soft Computing},
    keywords = {Fuzzy systems,Metaheuristic opti,Sparse regression,fuzzy systems,metaheuristic optimization,sparse regression},
    number = {20},
    pages = {15113--15127},
    publisher = {Springer Berlin Heidelberg},
    title = {{Approximation of two-variable functions using high-order Takagi-Sugeno fuzzy systems, sparse regressions, and metaheuristic optimization}},
    url = {https://doi.org/10.1007/s00500-020-05238-3},
    volume = {24},
    year = {2020}
    }
  • K. Wiktorowicz and T. Krzeszowski, „Training High-Order Takagi-Sugeno Fuzzy Systems Using Batch Least Squares and Particle Swarm Optimization,” International Journal of Fuzzy Systems, vol. 22, iss. 1, p. 22–34, 2020. doi:10.1007/s40815-019-00747-2
    [BibTeX] [Abstract] [Download PDF]
    This paper proposes two methods for training Takagi–Sugeno (T-S) fuzzy systems using batch least squares (BLS) and particle swarm optimization (PSO). The T-S system is considered with triangular and Gaussian membership functions in the antecedents and higher-order polynomials in the consequents of fuzzy rules. In the first method, the BLS determines the polynomials in a system in which the fuzzy sets are known. In the second method, the PSO algorithm determines the fuzzy sets, whereas the BLS determines the polynomials. In this paper, the ridge regression is used to stabilize the solution when the problem is close to the singularity. Thanks to this, the proposed methods can be applied when the number of observations is less than the number of predictors. Moreover, the leave-one-out cross-validation is used to avoid overfitting and this way to choose the structure of a fuzzy model. A method of obtaining piecewise linear regression by means of the zero-order T-S system is also presented.
    @Article{Wiktorowicz2020c,
    author="Wiktorowicz, Krzysztof
    and Krzeszowski, Tomasz",
    title="Training High-Order Takagi-Sugeno Fuzzy Systems Using Batch Least Squares and Particle Swarm Optimization",
    journal="International Journal of Fuzzy Systems",
    year="2020",
    month="Feb",
    day="01",
    volume="22",
    number="1",
    pages="22--34",
    abstract="This paper proposes two methods for training Takagi--Sugeno (T-S) fuzzy systems using batch least squares (BLS) and particle swarm optimization (PSO). The T-S system is considered with triangular and Gaussian membership functions in the antecedents and higher-order polynomials in the consequents of fuzzy rules. In the first method, the BLS determines the polynomials in a system in which the fuzzy sets are known. In the second method, the PSO algorithm determines the fuzzy sets, whereas the BLS determines the polynomials. In this paper, the ridge regression is used to stabilize the solution when the problem is close to the singularity. Thanks to this, the proposed methods can be applied when the number of observations is less than the number of predictors. Moreover, the leave-one-out cross-validation is used to avoid overfitting and this way to choose the structure of a fuzzy model. A method of obtaining piecewise linear regression by means of the zero-order T-S system is also presented.",
    issn="2199-3211",
    doi="10.1007/s40815-019-00747-2",
    url="https://doi.org/10.1007/s40815-019-00747-2"
    }

2019

  • J. Wubben, F. Fabra, C. T. Calafate, T. Krzeszowski, J. M. Marquez-Barja, J. C. Cano, and P. Manzoni, „Accurate landing of unmanned aerial vehicles using ground pattern recognition,” Electronics (Switzerland), vol. 8, iss. 12, p. 1532, 2019. doi:10.3390/electronics8121532
    [BibTeX] [Abstract]
    Over the last few years, several researchers have been developing protocols and applications in order to autonomously land unmanned aerial vehicles (UAVs). However, most of the proposed protocols rely on expensive equipment or do not satisfy the high precision needs of some UAV applications such as package retrieval and delivery or the compact landing of UAV swarms. Therefore, in this work, a solution for high precision landing based on the use of ArUco markers is presented. In the proposed solution, a UAV equipped with a low-cost camera is able to detect ArUco markers sized 56 56 cm from an altitude of up to 30 m. Once the marker is detected, the UAV changes its flight behavior in order to land on the exact position where the marker is located. The proposal was evaluated and validated using both the ArduSim simulation platform and real UAV flights. The results show an average offset of only 11 cm from the target position, which vastly improves the landing accuracy compared to the traditional GPS-based landing, which typically deviates from the intended target by 1 to 3 m.
    @article{Wubben2019,
    abstract = {Over the last few years, several researchers have been developing protocols and applications in order to autonomously land unmanned aerial vehicles (UAVs). However, most of the proposed protocols rely on expensive equipment or do not satisfy the high precision needs of some UAV applications such as package retrieval and delivery or the compact landing of UAV swarms. Therefore, in this work, a solution for high precision landing based on the use of ArUco markers is presented. In the proposed solution, a UAV equipped with a low-cost camera is able to detect ArUco markers sized 56  56 cm from an altitude of up to 30 m. Once the marker is detected, the UAV changes its flight behavior in order to land on the exact position where the marker is located. The proposal was evaluated and validated using both the ArduSim simulation platform and real UAV flights. The results show an average offset of only 11 cm from the target position, which vastly improves the landing accuracy compared to the traditional GPS-based landing, which typically deviates from the intended target by 1 to 3 m.},
    author = {Wubben, Jamie and Fabra, Francisco and Calafate, Carlos T. and Krzeszowski, Tomasz and Marquez-Barja, Johann M. and Cano, Juan Carlos and Manzoni, Pietro},
    doi = {10.3390/electronics8121532},
    issn = {20799292},
    journal = {Electronics (Switzerland)},
    keywords = {ArUco marker,ArduSim,Autonomous landing,UAV,Vision-based},
    number = {12},
    pages = {1532},
    title = {{Accurate landing of unmanned aerial vehicles using ground pattern recognition}},
    volume = {8},
    year = {2019}
    }
  • J. Wubben, F. Fabra, C. T. Calafate, T. Krzeszowski, J. M. Marquez-Barja, J. Cano, and P. Manzoni, „A vision-based system for autonomous vertical landing of unmanned aerial vehicles,” in 2019 IEEE/ACM 23rd International Symposium on Distributed Simulation and Real Time Applications (DS-RT), 2019, pp. 1-7. doi:10.1109/DS-RT47707.2019.8958701
    [BibTeX] [Abstract]
    Over the last few years, different researchers have been developing protocols and applications in order to land unmanned aerial vehicles (UAVs) autonomously. However, most of the proposed protocols rely on expensive equipment or do not satisfy the high precision needs of some UAV applications, such as package retrieval and delivery. Therefore, in this paper, we present a solution for high precision landing based on the use of ArUco markers. In our solution, a UAV equipped with a camera is able to detect ArUco markers from an altitude of 20 meters. Once the marker is detected, the UAV changes its flight behavior in order to land on the exact position where the marker is located. We evaluated our proposal using our own UAV simulation platform (ArduSim), and validated it using real UAVs. The results show an average offset of only 11 centimeters, which vastly improves the landing accuracy compared to the traditional GPS-based landing, that typically deviates from the intended target by 1 to 3 meters.
    @INPROCEEDINGS{Wubben2019DS-RT,
    author={J. {Wubben} and F. {Fabra} and C. T. {Calafate} and T. {Krzeszowski} and J. M. {Marquez-Barja} and J. {Cano} and P. {Manzoni}},
    booktitle={2019 IEEE/ACM 23rd International Symposium on Distributed Simulation and Real Time Applications (DS-RT)},
    title={A vision-based system for autonomous vertical landing of unmanned aerial vehicles},
    year={2019},
    volume={},
    number={},
    pages={1-7},
    keywords={},
    doi={10.1109/DS-RT47707.2019.8958701},
    abstract = {Over the last few years, different researchers have been developing protocols and applications in order to land unmanned aerial vehicles (UAVs) autonomously. However, most of the proposed protocols rely on expensive equipment or do not satisfy the high precision needs of some UAV applications, such as package retrieval and delivery. Therefore, in this paper, we present a solution for high precision landing based on the use of ArUco markers. In our solution, a UAV equipped with a camera is able to detect ArUco markers from an altitude of 20 meters. Once the marker is detected, the UAV changes its flight behavior in order to land on the exact position where the marker is located. We evaluated our proposal using our own UAV simulation platform (ArduSim), and validated it using real UAVs. The results show an average offset of only 11 centimeters, which vastly improves the landing accuracy compared to the traditional GPS-based landing, that typically deviates from the intended target by 1 to 3 meters.},
    ISSN={1550-6525},
    month={Oct},}
  • B. Kwolek, A. Michalczuk, T. Krzeszowski, A. Switonski, H. Josinski, and K. Wojciechowski, „Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition,” Multimedia Tools and Applications, vol. 78, iss. 22, p. 32437–32465, 2019. doi:10.1007/s11042-019-07945-y
    [BibTeX] [Abstract] [Download PDF]
    We introduce synchronized and calibrated multi-view video and motion capture dataset for motion analysis and gait identification. The 3D gait dataset consists of 166 data sequences with 32 people. In 128 data sequences, each of 32 individuals was dressed in his/her clothes, in 24 data sequences, 6 of 32 performers changed clothes, and in 14 data sequences, 7 of the performers had a backpack on his/her back. In a single recording session, every performer walked from right to left, then from left to right, and afterwards on the diagonal from upper-right to bottom-left and from bottom-left to upper-right corner of a rectangular scene. We demonstrate that a baseline algorithm achieves promising results in a challenging scenario, in which gallery/training data were collected in walks perpendicular/facing to the cameras, whereas the probe/testing data were collected in diagonal walks. We compare performances of biometric gait recognition that were achieved on marker-less and marker-based 3D data. We present recognition performances, which were achieved by a convolutional neural network and classic classifiers operating on gait signatures obtained by multilinear principal component analysis. The availability of synchronized multi-view image sequences with 3D locations of body markers creates a number of possibilities for extraction of discriminative gait signatures. The gait data are available at http://bytom.pja.edu.pl/projekty/hm-gpjatk/.
    @Article{Kwolek2019,
    author="Kwolek, Bogdan
    and Michalczuk, Agnieszka
    and Krzeszowski, Tomasz
    and Switonski, Adam
    and Josinski, Henryk
    and Wojciechowski, Konrad",
    title="Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition",
    journal="Multimedia Tools and Applications",
    year="2019",
    month="Nov",
    day="01",
    volume="78",
    number="22",
    pages="32437--32465",
    abstract="We introduce synchronized and calibrated multi-view video and motion capture dataset for motion analysis and gait identification. The 3D gait dataset consists of 166 data sequences with 32 people. In 128 data sequences, each of 32 individuals was dressed in his/her clothes, in 24 data sequences, 6 of 32 performers changed clothes, and in 14 data sequences, 7 of the performers had a backpack on his/her back. In a single recording session, every performer walked from right to left, then from left to right, and afterwards on the diagonal from upper-right to bottom-left and from bottom-left to upper-right corner of a rectangular scene. We demonstrate that a baseline algorithm achieves promising results in a challenging scenario, in which gallery/training data were collected in walks perpendicular/facing to the cameras, whereas the probe/testing data were collected in diagonal walks. We compare performances of biometric gait recognition that were achieved on marker-less and marker-based 3D data. We present recognition performances, which were achieved by a convolutional neural network and classic classifiers operating on gait signatures obtained by multilinear principal component analysis. The availability of synchronized multi-view image sequences with 3D locations of body markers creates a number of possibilities for extraction of discriminative gait signatures. The gait data are available at http://bytom.pja.edu.pl/projekty/hm-gpjatk/.",
    issn="1573-7721",
    doi="10.1007/s11042-019-07945-y",
    url="https://doi.org/10.1007/s11042-019-07945-y"
    }
  • K. Przednowek, T. Krzeszowski, J. Iskra, and K. Wiktorowicz, Wspomaganie procesu treningowego w biegach przez płotki z wykorzystaniem modelowania komputerowego, Uniwersytet Rzeszowski, 2019.
    [BibTeX]
    @Book{Przednowek2019mon,
    title = {Wspomaganie procesu treningowego w biegach przez p\l{}otki z wykorzystaniem modelowania komputerowego},
    publisher = {Uniwersytet Rzeszowski},
    year = {2019},
    author = {Przednowek, Krzysztof and Krzeszowski, Tomasz and Iskra, Janusz and Wiktorowicz, Krzysztof},
    }
  • T. Krzeszowski, K. Przednowek, K. Wiktorowicz, and J. Iskra, „The Application of Multiview Human Body Tracking on the Example of Hurdle Clearance,” in Sport Science Research and Technology Support, Cham, 2019, p. 116–127. doi:10.1007/978-3-030-14526-2_8
    [BibTeX] [Abstract] [Download PDF]
    This initial research presents the multiview human body tracking method as a tool to measure hurdle clearance parameters. This study was conducted on high level hurdlers, who were members of the Polish national team. The video sequences were recorded by a multicamera system consisting of three 100Hz Full HD cameras. The sequences were registered under the simulated starting conditions of a 110m hurdles race. Kinematic parameters were estimated based on the analysis of images from the multicamera system. These parameters were compared with the parameters obtained from ground truth poses. Mean absolute error and mean relative error were selected as the quality criteria. The main advantage of the method presented here is that it does not need any special clothes, markers or the support of other estimation techniques.
    @InProceedings{KrzeszowskiSSRTS2019,
    author="Krzeszowski, Tomasz
    and Przednowek, Krzysztof
    and Wiktorowicz, Krzysztof
    and Iskra, Janusz",
    editor="Cabri, Jan
    and Pezarat-Correia, Pedro
    and Vilas-Boas, Jo{\~a}o",
    title="The Application of Multiview Human Body Tracking on the Example of Hurdle Clearance",
    booktitle="Sport Science Research and Technology Support",
    year="2019",
    publisher="Springer International Publishing",
    volume={975},
    address="Cham",
    pages="116--127",
    abstract="This initial research presents the multiview human body tracking method as a tool to measure hurdle clearance parameters. This study was conducted on high level hurdlers, who were members of the Polish national team. The video sequences were recorded by a multicamera system consisting of three 100Hz Full HD cameras. The sequences were registered under the simulated starting conditions of a 110m hurdles race. Kinematic parameters were estimated based on the analysis of images from the multicamera system. These parameters were compared with the parameters obtained from ground truth poses. Mean absolute error and mean relative error were selected as the quality criteria. The main advantage of the method presented here is that it does not need any special clothes, markers or the support of other estimation techniques.",
    isbn="978-3-030-14526-2",
    doi = {10.1007/978-3-030-14526-2_8},
    url = {https://link.springer.com/chapter/10.1007/978-3-030-14526-2_8}
    }
  • K. Przednowek, K. Wiktorowicz, T. Krzeszowski, and J. Iskra, „A web-oriented expert system for planning hurdles race training programmes,” Neural Computing and Applications, vol. 31, iss. 11, p. 7227–7243, 2019. doi:10.1007/s00521-018-3559-1
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a web-oriented expert system, named iHurdling, to predict results and generate training loads for 110 and 400m hurdles races. The database contains 40 annual training programmes for the 110m hurdles and 48 programmes for the 400m hurdles. The predictive models include linear regressions in the form of ordinary least squares, ridge, LASSO, elastic net and nonlinear models in the form of a radial basis function neural network and fuzzy rule-based system. The leave-one-out cross-validation method is used to compare, and choose the best model. It shows that the proposed fuzzy-based model has the lowest validation error. The developed web application can support a coach in planning training programmes for hurdles races. It allows the athlete’s results to be predicted and can generate training loads for an athlete, selected from database. The application can be run on a computer or a mobile device. The system was implemented using the R programming language with the Shiny framework and additional packages. The limitations of the presented approach are related to the lack of consideration of an athlete’s physiological and psychological parameters, but the generated training programs might be used as a suggestion for the coach.
    @Article{Przednowek2019,
    author="Przednowek, Krzysztof
    and Wiktorowicz, Krzysztof
    and Krzeszowski, Tomasz
    and Iskra, Janusz",
    title="A web-oriented expert system for planning hurdles race training programmes",
    journal="Neural Computing and Applications",
    year="2019",
    month="Nov",
    day="01",
    volume="31",
    number="11",
    pages="7227--7243",
    abstract="This paper presents a web-oriented expert system, named iHurdling, to predict results and generate training loads for 110 and 400m hurdles races. The database contains 40 annual training programmes for the 110m hurdles and 48 programmes for the 400m hurdles. The predictive models include linear regressions in the form of ordinary least squares, ridge, LASSO, elastic net and nonlinear models in the form of a radial basis function neural network and fuzzy rule-based system. The leave-one-out cross-validation method is used to compare, and choose the best model. It shows that the proposed fuzzy-based model has the lowest validation error. The developed web application can support a coach in planning training programmes for hurdles races. It allows the athlete's results to be predicted and can generate training loads for an athlete, selected from database. The application can be run on a computer or a mobile device. The system was implemented using the R programming language with the Shiny framework and additional packages. The limitations of the presented approach are related to the lack of consideration of an athlete's physiological and psychological parameters, but the generated training programs might be used as a suggestion for the coach.",
    issn="1433-3058",
    doi="10.1007/s00521-018-3559-1",
    url="https://doi.org/10.1007/s00521-018-3559-1"
    }

2018

  • A. Switonski, T. Krzeszowski, H. Josinski, B. Kwolek, and K. Wojciechowski, „Gait recognition on the basis of markerless motion tracking and DTW transform,” IET Biometrics, vol. 7, iss. 5, p. 415–422, 2018. doi:10.1049/iet-bmt.2017.0134
    [BibTeX] [Abstract] [Download PDF]
    In this study, a framework for view-invariant gait recognition on the basis of markerless motion tracking and dynamic time warping (DTW) transform is presented. The system consists of a proposed markerless motion capture system as well as introduced classification method of mocap data. The markerless system estimates the three-dimensional locations of skeleton driven joints. Such skeleton-driven point clouds represent poses over time. The authors align point clouds in every pair of frames by calculating the minimal sum of squared distances between the corresponding joints. A point cloud distance measure with temporal context has been utilised in k-nearest neighbours algorithm to compare time instants of motion sequences. To enhance the generalisation of the recognition and to shorten the processing time, for every individual a single multidimensional time series among several multidimensional time series describing the individual’s gait is established. The correct classification rate has been determined on the basis of a real dataset of human gait. It contains 230 gait cycles of 22 subjects. The tracking results on the basis of markerless motion capture are referenced to Vicon system, whereas the achieved accuracies of recognition are compared with the ones obtained by DTW that is based on rotational data.
    @article{Switonski2018,
    abstract = {In this study, a framework for view-invariant gait recognition on the basis of markerless motion tracking and dynamic time warping (DTW) transform is presented. The system consists of a proposed markerless motion capture system as well as introduced classification method of mocap data. The markerless system estimates the three-dimensional locations of skeleton driven joints. Such skeleton-driven point clouds represent poses over time. The authors align point clouds in every pair of frames by calculating the minimal sum of squared distances between the corresponding joints. A point cloud distance measure with temporal context has been utilised in k-nearest neighbours algorithm to compare time instants of motion sequences. To enhance the generalisation of the recognition and to shorten the processing time, for every individual a single multidimensional time series among several multidimensional time series describing the individual's gait is established. The correct classification rate has been determined on the basis of a real dataset of human gait. It contains 230 gait cycles of 22 subjects. The tracking results on the basis of markerless motion capture are referenced to Vicon system, whereas the achieved accuracies of recognition are compared with the ones obtained by DTW that is based on rotational data.},
    author = {Switonski, Adam and Krzeszowski, Tomasz and Josinski, Henryk and Kwolek, Bogdan and Wojciechowski, Konrad},
    doi = {10.1049/iet-bmt.2017.0134},
    issn = {2047-4938},
    journal = {IET Biometrics},
    number = {5},
    pages = {415--422},
    title = {{Gait recognition on the basis of markerless motion tracking and DTW transform}},
    url = {http://digital-library.theiet.org/content/journals/10.1049/iet-bmt.2017.0134},
    volume = {7},
    year = {2018},
    note = {IF2018: 2.092}
    }
  • B. Rymut, T. Krzeszowski, K. Przednowek, K. H. Przednowek, and J. Iskra, „Kinematic Analysis of Hurdle Clearance using a Mobile Device,” in Proceedings of the 6th International Congress on Sport Sciences Research and Technology Support – Volume 1 (icSPORTS 2018), Seville, Spain, 2018, p. 49–55. doi:10.5220/0006933600490055
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a human motion tracking method using a mobile device. The proposed method may be used as a tool to measure hurdle clearance kinematic parameters and help coaches to evaluate the athlete’s technique. The video recordings were made under simulated starting conditions of a 100 m women hurdle race. Kinematic parameters were estimated based on an analysis of images sequence from a mobile device. The images were recorded on a HTC M8s smartphone with a resolution of 1920×1080 pixels and with a frequency of 30 Hz. The system was tested on two mobile development platforms and three image sequences. The proposed method does not use any markers, special clothes or other estimation support techniques. The analysis conducted showed that the smallest errors were calculated for the height of centre of mass, while the biggest errors were observed for the bending angle of the knee of the trail leg.
    @inproceedings{Rymut2018,
    abstract = {This paper presents a human motion tracking method using a mobile device. The proposed method may be used as a tool to measure hurdle clearance kinematic parameters and help coaches to evaluate the athlete's technique. The video recordings were made under simulated starting conditions of a 100 m women hurdle race. Kinematic parameters were estimated based on an analysis of images sequence from a mobile device. The images were recorded on a HTC M8s smartphone with a resolution of 1920x1080 pixels and with a frequency of 30 Hz. The system was tested on two mobile development platforms and three image sequences. The proposed method does not use any markers, special clothes or other estimation support techniques. The analysis conducted showed that the smallest errors were calculated for the height of centre of mass, while the biggest errors were observed for the bending angle of the knee of the trail leg.},
    address = {Seville, Spain},
    author = {Rymut, Boguslaw and Krzeszowski, Tomasz and Przednowek, Krzysztof and Przednowek, Karolina H and Iskra, Janusz},
    booktitle = {Proceedings of the 6th International Congress on Sport Sciences Research and Technology Support - Volume 1 (icSPORTS 2018)},
    doi = {10.5220/0006933600490055},
    isbn = {978-989-758-325-4},
    pages = {49--55},
    publisher = {Scitepress},
    title = {{Kinematic Analysis of Hurdle Clearance using a Mobile Device}},
    url = {http://dx.doi.org/10.5220/0006933600490055},
    year = {2018}
    }
  • T. Krzeszowski, K. Wiktorowicz, and K. Przednowek, „Comparison of selected fuzzy PSO algorithms,” in Recent Advances in Computational Optimization, Studies in Computational Intelligence, 2018, p. 107–122. doi:10.1007/978-3-319-59861-1_7
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a comparison of selected fuzzy particle swarm optimization algorithms. Two non-fuzzy and four fuzzy algorithms are considered. The Takagi–Sugeno fuzzy system is used to change the parameters of these algorithms. A modified fuzzy particle swarm optimization method is proposed in which each of the particles has its own inertia weight and coefficients of the cognitive and social components. The evaluation is based on the common nonlinear benchmark functions frequently used for testing optimization methods. The ratings of the algorithms are assigned on the basis of the mean of the objective function final value.
    @inproceedings{Krzeszowski2018,
    abstract = {This paper presents a comparison of selected fuzzy particle swarm optimization algorithms. Two non-fuzzy and four fuzzy algorithms are considered. The Takagi–Sugeno fuzzy system is used to change the parameters of these algorithms. A modified fuzzy particle swarm optimization method is proposed in which each of the particles has its own inertia weight and coefficients of the cognitive and social components. The evaluation is based on the common nonlinear benchmark functions frequently used for testing optimization methods. The ratings of the algorithms are assigned on the basis of the mean of the objective function final value.},
    author = {Krzeszowski, Tomasz and Wiktorowicz, Krzysztof and Przednowek, Krzysztof},
    booktitle = {Recent Advances in Computational Optimization, Studies in Computational Intelligence},
    doi = {10.1007/978-3-319-59861-1_7},
    isbn = {978-3-319-59860-4},
    pages = {107--122},
    publisher = {Springer Cham},
    title = {{Comparison of selected fuzzy PSO algorithms}},
    url = {https://link.springer.com/chapter/10.1007/978-3-319-59861-1_7},
    volume = {717},
    year = {2018}
    }
  • K. Przednowek, K. Wiktorowicz, T. Krzeszowski, M. Tumidajewicz, and J. Iskra, „Mobile application for predictive modelling in hurdles race,” in 2018 2nd International Conference on Technology and Innovation in Sports, Health and Wellbeing (TISHW), Thessaloniki, Greece, 2018, p. 1–7. doi:10.1109/TISHW.2018.8559575
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a mobile expert system for Android platform, named R-tificial Trainer, to support the work of a hurdles coach in planning training programmes. The main feature of the developed application is the ability to generate training loads and predict results for an athlete. It includes a database of players and allows the user to generate training plan in PDF format. The application has been tested on a dataset of athletes practising hurdles on the 110 metres. The database contains 120 training programmes made by 18 athletes. The application uses the Predictive Model Markup Language standard. The predictive models include linear models in the form of ordinary least squares and LASSO regressions and nonlinear model in the form of a multilayer perceptron with exponential function. To choose the best method, the leave-one-out cross-validation is used. The lowest validation error was achieved by multilayer perceptron.
    @inproceedings{Przednowek2018b,
    abstract = {This paper presents a mobile expert system for Android platform, named R-tificial Trainer, to support the work of a hurdles coach in planning training programmes. The main feature of the developed application is the ability to generate training loads and predict results for an athlete. It includes a database of players and allows the user to generate training plan in PDF format. The application has been tested on a dataset of athletes practising hurdles on the 110 metres. The database contains 120 training programmes made by 18 athletes. The application uses the Predictive Model Markup Language standard. The predictive models include linear models in the form of ordinary least squares and LASSO regressions and nonlinear model in the form of a multilayer perceptron with exponential function. To choose the best method, the leave-one-out cross-validation is used. The lowest validation error was achieved by multilayer perceptron.},
    address = {Thessaloniki, Greece},
    author = {Przednowek, K and Wiktorowicz, K and Krzeszowski, T and Tumidajewicz, M and Iskra, J},
    booktitle = {2018 2nd International Conference on Technology and Innovation in Sports, Health and Wellbeing (TISHW)},
    doi = {10.1109/TISHW.2018.8559575},
    keywords = {Predictive Model Markup Language,expert systems,hurdles race,mobile application,predictive models},
    pages = {1--7},
    publisher = {IEEE},
    title = {{Mobile application for predictive modelling in hurdles race}},
    url = {https://doi.org/10.1109/TISHW.2018.8559575},
    year = {2018}
    }
  • K. Przednowek, T. Krzeszowski, K. Przednowek, and P. Lenik, „A System for Analysing the Basketball Free Throw Trajectory Based on Particle Swarm Optimization,” Applied Sciences, vol. 8, iss. 11, p. 2090, 2018. doi:10.3390/app8112090
    [BibTeX] [Abstract] [Download PDF]
    This paper describes a system for the automatic detection and tracking of a ball trajectory during a free throw. The tracking method is based on a particle swarm optimization (PSO) algorithm. The proposed method allows for the measurement of selected parameters of a basketball free throw trajectory. Ten parameters (four distances, three velocities, and three angle parameters) were taken into account. The research material included 200 sequences captured by a 100 Hz monocular camera. The study was based on a group of 30 basketball players who played in the Polish Second Division during the 2015/2016 season and the Youth Polish National Team in 2017. The experimental results showed the differences between the parameters in both missed and hit throws. The proposed system may be used in the training process as a tool to improve the technique of the free throw in basketball.
    @article{Przednowek2018,
    abstract = {This paper describes a system for the automatic detection and tracking of a ball trajectory during a free throw. The tracking method is based on a particle swarm optimization (PSO) algorithm. The proposed method allows for the measurement of selected parameters of a basketball free throw trajectory. Ten parameters (four distances, three velocities, and three angle parameters) were taken into account. The research material included 200 sequences captured by a 100 Hz monocular camera. The study was based on a group of 30 basketball players who played in the Polish Second Division during the 2015/2016 season and the Youth Polish National Team in 2017. The experimental results showed the differences between the parameters in both missed and hit throws. The proposed system may be used in the training process as a tool to improve the technique of the free throw in basketball.},
    author = {Przednowek, Krzysztof and Krzeszowski, Tomasz and Przednowek, Karolina and Lenik, Pawel},
    doi = {10.3390/app8112090},
    issn = {2076-3417},
    journal = {Applied Sciences},
    month = {oct},
    number = {11},
    pages = {2090},
    title = {{A System for Analysing the Basketball Free Throw Trajectory Based on Particle Swarm Optimization}},
    url = {https://doi.org/10.3390/app8112090},
    volume = {8},
    year = {2018}
    }

2017

  • J. Iskra, K. Przednowek, K. Wiktorowicz, and T. Krzeszowski, „The Use of Artificial Neural Networks in Supporting the Annual Training in 400 meter Hurdles,” Central European Journal of Sport Sciences and Medicine, vol. 17, iss. 1, p. 15–24, 2017. doi:10.18276/cej.2017.1-02
    [BibTeX] [Abstract] [Download PDF]
    This paper presents an evaluation of the annual cycle for 400 m hurdles using artificial neural networks. The analysis included 21 Polish national team hurdlers. In planning the annual cycle, 27 variables were used, where 5 variables describe the competitor and 22 variables represent the training loads. In the presented solution, the task of generating training loads for the assumed result were considered. The neural models were evaluated by cross-validation method. The smallest error was obtained for the radial basis function network with nine neurons in the hidden layer. The performed analysis shows that at each phase of training the structure of training loads is different.
    @article{Iskra2017,
    abstract = {This paper presents an evaluation of the annual cycle for 400 m hurdles using artificial neural networks. The analysis included 21 Polish national team hurdlers. In planning the annual cycle, 27 variables were used, where 5 variables describe the competitor and 22 variables represent the training loads. In the presented solution, the task of generating training loads for the assumed result were considered. The neural models were evaluated by cross-validation method. The smallest error was obtained for the radial basis function network with nine neurons in the hidden layer. The performed analysis shows that at each phase of training the structure of training loads is different.
    },
    author = {Iskra, Janusz and Przednowek, Krzysztof and Wiktorowicz, Krzysztof and Krzeszowski, Tomasz},
    doi = {10.18276/cej.2017.1-02},
    issn = {2300-9705},
    journal = {Central European Journal of Sport Sciences and Medicine},
    number = {1},
    pages = {15--24},
    title = {{The Use of Artificial Neural Networks in Supporting the Annual Training in 400 meter Hurdles}},
    url = {https://wnus.edu.pl/cejssm/en/issue/451/article/7645/},
    volume = {17},
    year = {2017}
    }
  • K. Przednowek, J. Iskra, K. Wiktorowicz, T. Krzeszowski, and A. Maszczyk, „Planning Training Loads for the 400 M Hurdles in Three-Month Mesocycles Using Artificial Neural Networks,” Journal of Human Kinetics, vol. 60, iss. 1, p. 175–189, 2017. doi:10.1515/hukin-2017-0101
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a novel approach to planning training loads in hurdling using artificial neural networks. The neural models performed the task of generating loads for athletes' training for the 400 meters hurdles. All the models were calculated based on the training data of 21 Polish National Team hurdlers, aged 22.25 ± 1.96, competing between 1989 and 2012. The analysis included 144 training plans that represented different stages in the annual training cycle. The main contribution of this paper is to develop neural models for planning training loads for the entire career of a typical hurdler. In the models, 29 variables were used, where four characterized the runner and 25 described the training process. Two artificial neural networks were used: a multi-layer perceptron and a network with radial basis functions. To assess the quality of the models, the leave-one-out cross-validation method was used in which the Normalized Root Mean Squared Error was calculated. The analysis shows that the method generating the smallest error was the radial basis function network with nine neurons in the hidden layer. Most of the calculated training loads demonstrated a non-linear relationship across the entire competitive period. The resulting model can be used as a tool to assist a coach in planning training loads during a selected training period.
    @article{Przednowek2017,
    abstract = {This paper presents a novel approach to planning training loads in hurdling using artificial neural networks. The neural models performed the task of generating loads for athletes' training for the 400 meters hurdles. All the models were calculated based on the training data of 21 Polish National Team hurdlers, aged 22.25 ± 1.96, competing between 1989 and 2012. The analysis included 144 training plans that represented different stages in the annual training cycle. The main contribution of this paper is to develop neural models for planning training loads for the entire career of a typical hurdler. In the models, 29 variables were used, where four characterized the runner and 25 described the training process. Two artificial neural networks were used: a multi-layer perceptron and a network with radial basis functions. To assess the quality of the models, the leave-one-out cross-validation method was used in which the Normalized Root Mean Squared Error was calculated. The analysis shows that the method generating the smallest error was the radial basis function network with nine neurons in the hidden layer. Most of the calculated training loads demonstrated a non-linear relationship across the entire competitive period. The resulting model can be used as a tool to assist a coach in planning training loads during a selected training period.},
    author = {Przednowek, Krzysztof and Iskra, Janusz and Wiktorowicz, Krzysztof and Krzeszowski, Tomasz and Maszczyk, Adam},
    doi = {10.1515/hukin-2017-0101},
    issn = {18997562},
    journal = {Journal of Human Kinetics},
    keywords = {400 m hurdles,artificial neural network,training loads},
    number = {1},
    pages = {175--189},
    pmid = {29339998},
    title = {{Planning Training Loads for the 400 M Hurdles in Three-Month Mesocycles Using Artificial Neural Networks}},
    url = {https://content.sciendo.com/view/journals/hukin/60/1/article-p175.xml},
    volume = {60},
    year = {2017}
    }
  • T. Krzeszowski, K. Przednowek, K. Wiktorowicz, and J. Iskra, „Multiview Human Body Tracking of Hurdle Clearance: A Case Study,” in Proceedings of the 5th International Congress on Sport Sciences Research and Technology Support, Funchal, Madeira, Portugal, 2017, p. 83–88. doi:10.5220/0006498400830088
    [BibTeX] [Abstract] [Download PDF]
    This initial research is a case study that uses a multiview human body tracking method as a tool to measure hurdle clearance kinematic parameters. This study is conducted on a hurdler representing a high sport level, who is a participant in the European and World Championships and the Olympic Games. The video recordings were made under simulated starting conditions of a 110 m hurdle race. Kinematic parameters are estimated based on the analysis of images from a multicamera system. The images were recorded with a resolution of 1920×1080 and with a frequency of 100 Hz. The proposed method does not use any special clothes, markers or other estimation support techniques. The parameters of the hurdle clearance were compared with the parameters obtained from ground truth poses. Mean Absolute Error and Mean Relative Error were used as the quality criteria.
    @inproceedings{Krzeszowski2017,
    abstract = {This initial research is a case study that uses a multiview human body tracking method as a tool to measure hurdle clearance kinematic parameters. This study is conducted on a hurdler representing a high sport level, who is a participant in the European and World Championships and the Olympic Games. The video recordings were made under simulated starting conditions of a 110 m hurdle race. Kinematic parameters are estimated based on the analysis of images from a multicamera system. The images were recorded with a resolution of 1920x1080 and with a frequency of 100 Hz. The proposed method does not use any special clothes, markers or other estimation support techniques. The parameters of the hurdle clearance were compared with the parameters obtained from ground truth poses. Mean Absolute Error and Mean Relative Error were used as the quality criteria.},
    address = {Funchal, Madeira, Portugal},
    author = {Krzeszowski, Tomasz and Przednowek, Krzysztof and Wiktorowicz, Krzysztof and Iskra, Janusz},
    booktitle = {Proceedings of the 5th International Congress on Sport Sciences Research and Technology Support},
    doi = {10.5220/0006498400830088},
    isbn = {978-989-758-269-1},
    keywords = {110 m Hurdle Race,Human Motion Tracking,Multicamera System,Particle Swarm Optimisation},
    pages = {83--88},
    publisher = {Scitepress},
    title = {{Multiview Human Body Tracking of Hurdle Clearance: A Case Study}},
    url = {http://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0006498400830088},
    year = {2017}
    }
  • K. Przednowek, J. Iskra, T. Krzeszowski, and K. H. Przednowek, „Application of Artificial Neural Models for Planning Sport Training in 110m Hurdles,” in Proceedings of the 5th International Congress on Neurotechnology, Electronics and Informatics, Funchal, Madeira, Portugal, 2017, p. 41–46. doi:10.5220/0006499400410046
    [BibTeX] [Abstract] [Download PDF]
    This paper presents the use of artificial neural networks for planning sport training in 110 meters hurdles. The model was calculated based on the training data of the Polish National Team hurdlers. The analysis was based on 120 training plans that represent a different period in the annual training cycle. The MLP and RBF networks were used in this study as a predictive model. The neural network developed has 6 inputs representing the parameters of the athlete and 15 outputs representing the training loads. To evaluate the model, a set of 20 records were used. The smallest prediction error was obtained for the multilayer perceptron with 9 neurons in the hidden layer and a hyperbolic tangent as the activation functions. The resulting model may be used as a tool to assist coaches in planning training loads during the selected training period.
    @inproceedings{Przednowek2017a,
    abstract = {This paper presents the use of artificial neural networks for planning sport training in 110 meters hurdles. The model was calculated based on the training data of the Polish National Team hurdlers. The analysis was based on 120 training plans that represent a different period in the annual training cycle. The MLP and RBF networks were used in this study as a predictive model. The neural network developed has 6 inputs representing the parameters of the athlete and 15 outputs representing the training loads. To evaluate the model, a set of 20 records were used. The smallest prediction error was obtained for the multilayer perceptron with 9 neurons in the hidden layer and a hyperbolic tangent as the activation functions. The resulting model may be used as a tool to assist coaches in planning training loads during the selected training period.},
    address = {Funchal, Madeira, Portugal},
    author = {Przednowek, Krzysztof and Iskra, Janusz and Krzeszowski, Tomasz and Przednowek, Karolina H},
    booktitle = {Proceedings of the 5th International Congress on Neurotechnology, Electronics and Informatics},
    doi = {10.5220/0006499400410046},
    isbn = {9789897582707},
    keywords = {110m Hurdle Race,Artificial Neural Networks,Sport Training},
    pages = {41--46},
    publisher = {Scitepress},
    title = {{Application of Artificial Neural Models for Planning Sport Training in 110m Hurdles}},
    url = {http://www.scitepress.org/PublicationsDetail.aspx?ID=YNVIq8UwnRE=&t=1},
    year = {2017}
    }
  • J. Iskra, K. Przednowek, T. Krzeszowski, K. Wiktorowicz, and M. Pietrzak, „Kinematic Analysis of the Upper Limbs in Stepping over the Hurdle – The Use of IMU-based Motion Capture,” in Proceedings of the 5th International Congress on Sport Sciences Research and Technology Support, Funchal, Madeira, Portugal, 2017, p. 102–106. doi:10.5220/0006503101020106
    [BibTeX] [Abstract] [Download PDF]
    This paper presents an analysis of the kinematic parameters of the upper limbs in stepping over the hurdle. Stepping over the hurdle is a specific exercise practised throughout the year. In this exercise, three key points were analysed in take-off, flight and landing phases. The aim of the study was to use the IMU-based (inertial measurement unit) motion capture system to evaluate the movement of the hurdlers' upper limbs while stepping over the hurdle using both the better leg, and the worse leg. The sequences were obtained using 18 sensors working at a frequency of 120 Hz. The analysis was made using two high-achieving athletes. This paper presents the linear velocities and the trajectory of selected segments of the upper limbs. In most cases the velocities of the segments were higher for the better leg. The analysis shows that during the specific exercise of stepping over the hurdle attention should be paid to the movement of the trail arm in the landing phase.
    @inproceedings{Iskra2017a,
    abstract = {This paper presents an analysis of the kinematic parameters of the upper limbs in stepping over the hurdle. Stepping over the hurdle is a specific exercise practised throughout the year. In this exercise, three key points were analysed in take-off, flight and landing phases. The aim of the study was to use the IMU-based (inertial measurement unit) motion capture system to evaluate the movement of the hurdlers' upper limbs while stepping over the hurdle using both the better leg, and the worse leg. The sequences were obtained using 18 sensors working at a frequency of 120 Hz. The analysis was made using two high-achieving athletes. This paper presents the linear velocities and the trajectory of selected segments of the upper limbs. In most cases the velocities of the segments were higher for the better leg. The analysis shows that during the specific exercise of stepping over the hurdle attention should be paid to the movement of the trail arm in the landing phase.},
    address = {Funchal, Madeira, Portugal},
    author = {Iskra, Janusz and Przednowek, Krzysztof and Krzeszowski, Tomasz and Wiktorowicz, Krzysztof and Pietrzak, Michal},
    booktitle = {Proceedings of the 5th International Congress on Sport Sciences Research and Technology Support},
    doi = {10.5220/0006503101020106},
    isbn = {9789897582691},
    keywords = {Hurdling,IMU-based Motion Capture,Kinematic Parameters,Upper Limbs Movement},
    pages = {102--106},
    publisher = {Scitepress},
    title = {{Kinematic Analysis of the Upper Limbs in Stepping over the Hurdle - The Use of IMU-based Motion Capture}},
    url = {http://www.scitepress.org/PublicationsDetail.aspx?ID=F018sAN2qE4=&t=1},
    year = {2017}
    }

2016

  • T. Krzeszowski, K. Przednowek, K. Wiktorowicz, and J. Iskra, „Estimation of hurdle clearance parameters using a monocular human motion tracking method,” Computer Methods in Biomechanics and Biomedical Engineering, vol. 19, iss. 12, p. 1319–1329, 2016. doi:10.1080/10255842.2016.1139092
    [BibTeX] [Abstract] [Download PDF]
    This paper presents a method of monocular human motion tracking for estimation of hurdle clearance kinematic parameters. The analysis involved 10 image sequences of five hurdlers at various training levels. Recording of the sequences was carried out under simulated starting conditions of a 110 m hurdle race. The parameters were estimated using the particle swarm optimization algorithm and they are based on analysis of the images recorded with a 100 Hz camera. The proposed method does not involve using any special clothes, markers, inertial sensors, etc. As the quality criteria, the mean absolute error and mean relative error were used. The level of computed errors justifies the use of this method to estimate hurdle clearance parameters.
    @article{Krzeszowski2016,
    abstract = {This paper presents a method of monocular human motion tracking for estimation of hurdle clearance kinematic parameters. The analysis involved 10 image sequences of five hurdlers at various training levels. Recording of the sequences was carried out under simulated starting conditions of a 110 m hurdle race. The parameters were estimated using the particle swarm optimization algorithm and they are based on analysis of the images recorded with a 100 Hz camera. The proposed method does not involve using any special clothes, markers, inertial sensors, etc. As the quality criteria, the mean absolute error and mean relative error were used. The level of computed errors justifies the use of this method to estimate hurdle clearance parameters.},
    author = {Krzeszowski, Tomasz and Przednowek, Krzysztof and Wiktorowicz, Krzysztof and Iskra, Janusz},
    doi = {10.1080/10255842.2016.1139092},
    journal = {Computer Methods in Biomechanics and Biomedical Engineering},
    keywords = {Hurdle clearance,human motion tracking,kinematics parameters,monocular tracking,particle swarm optimization},
    number = {12},
    pages = {1319--1329},
    title = {{Estimation of hurdle clearance parameters using a monocular human motion tracking method}},
    url = {http://www.tandfonline.com/doi/full/10.1080/10255842.2016.1139092},
    volume = {19},
    year = {2016}
    }
  • T. Krzeszowski and K. Wiktorowicz, „Evaluation of selected fuzzy particle swarm optimization algorithms,” in Proceedings of the 2016 Federated Conference on Computer Science and Information Systems, FedCSIS 2016, Gdansk, Poland, 2016. doi:10.15439/2016F206
    [BibTeX] [Abstract] [Download PDF]
    This paper is devoted to an evaluation of selected fuzzy particle swarm optimization algorithms. Two non-fuzzy and four fuzzy algorithms are considered. The Takagi-Sugeno fuzzy system is utilized to change the parameters of these algorithms. A modified fuzzy particle swarm optimization method is proposed, in which each of the particles has its own inertia weight and coefficients of the cognitive and social components. The evaluation is based on the common nonlinear benchmark functions used for testing optimization methods. The ratings of the algorithms are assigned on the basis of the mean of the objective function and the relative success.
    @inproceedings{Krzeszowski2016a,
    abstract = {This paper is devoted to an evaluation of selected fuzzy particle swarm optimization algorithms. Two non-fuzzy and four fuzzy algorithms are considered. The Takagi-Sugeno fuzzy system is utilized to change the parameters of these algorithms. A modified fuzzy particle swarm optimization method is proposed, in which each of the particles has its own inertia weight and coefficients of the cognitive and social components. The evaluation is based on the common nonlinear benchmark functions used for testing optimization methods. The ratings of the algorithms are assigned on the basis of the mean of the objective function and the relative success.},
    address = {Gdansk, Poland},
    author = {Krzeszowski, T. and Wiktorowicz, K.},
    booktitle = {Proceedings of the 2016 Federated Conference on Computer Science and Information Systems, FedCSIS 2016},
    doi = {10.15439/2016F206},
    isbn = {9788360810903},
    publisher = {IEEE},
    title = {{Evaluation of selected fuzzy particle swarm optimization algorithms}},
    url = {https://ieeexplore.ieee.org/document/7733295},
    year = {2016}
    }
  • K. Przednowek, K. Wiktorowicz, T. Krzeszowski, and J. Iskra, „A fuzzy-based software tool used to predict 110m hurdles results during the annual training cycle,” in Proceedings of the 4th International Congress on Sport Sciences Research and Technology Support (icSPORTS 2016), Porto, Portugal, 2016, p. 176–181. doi:10.5220/0006043701760181
    [BibTeX] [Abstract] [Download PDF]
    This paper describes a fuzzy-based software tool for predicting results in the 110m hurdles. The predictive models were built on using 40 annual training cycles completed by 18 athletes. These models include: ordinary least squares regression, ridge regression, LASSO regression, elastic net regression and nonlinear fuzzy correction of least squares regression. In order to compare them, and choose the best model, leave-one-out cross-validation was used. This showed that the fuzzy corrector proposed in this paper has the lowest prediction error. The developed software can support a coach in planning an athlete’s annual training cycle. It allows the athlete’s results to be predicted, and in this way, for the best training loads to be selected. The tool is a web-based interactive application that can be run from a computer or a mobile device. The whole system was implemented using the R programming language with additional packages.
    @inproceedings{Przednowek2016,
    abstract = {This paper describes a fuzzy-based software tool for predicting results in the 110m hurdles. The predictive models were built on using 40 annual training cycles completed by 18 athletes. These models include: ordinary least squares regression, ridge regression, LASSO regression, elastic net regression and nonlinear fuzzy correction of least squares regression. In order to compare them, and choose the best model, leave-one-out cross-validation was used. This showed that the fuzzy corrector proposed in this paper has the lowest prediction error. The developed software can support a coach in planning an athlete's annual training cycle. It allows the athlete's results to be predicted, and in this way, for the best training loads to be selected. The tool is a web-based interactive application that can be run from a computer or a mobile device. The whole system was implemented using the R programming language with additional packages.},
    address = {Porto, Portugal},
    author = {Przednowek, Krzysztof and Wiktorowicz, Krzysztof and Krzeszowski, Tomasz and Iskra, Janusz},
    booktitle = {Proceedings of the 4th International Congress on Sport Sciences Research and Technology Support (icSPORTS 2016)},
    doi = {10.5220/0006043701760181},
    editor = {Correia, Pedro Pezarat and Cabri, Jan},
    isbn = {9789897582059},
    keywords = {110m hurdles,Fuzzy systems,Predictive models,R programming language},
    pages = {176--181},
    publisher = {Scitepress},
    title = {{A fuzzy-based software tool used to predict 110m hurdles results during the annual training cycle}},
    url = {http://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0006043701760181},
    year = {2016}
    }
  • K. Przednowek, J. Iskra, T. Krzeszowski, and K. Wiktorowicz, „Evaluation of Kinematic Parameters of Hurdles Clearance During Fatigue in Men’s 400m Hurdles — Research Using the Method of Computer Vision,” in Current research in motor control V: bridging motor control and biomechanics, Katowice, 2016, p. 232–238.
    [BibTeX]
    @inproceedings{Iskra2016,
    address = {Katowice},
    author = {Przednowek, Krzysztof and Iskra, Janusz and Krzeszowski, Tomasz and Wiktorowicz, Krzysztof},
    booktitle = {Current research in motor control V: bridging motor control and biomechanics},
    editor = {Slomka, Kajetan J. and Juras, Grzegorz},
    pages = {232--238},
    publisher = {AWF Katowice},
    title = {{Evaluation of Kinematic Parameters of Hurdles Clearance During Fatigue in Men's 400m Hurdles — Research Using the Method of Computer Vision}},
    year = {2016}
    }

2015

  • T. Krzeszowski, K. Przednowek, J. Iskra, and K. Wiktorowicz, „Monocular Tracking of Human Motion in Evaluation of Hurdle Clearance,” in Sports Science Research and Technology Support. icSPORTS 2014. Communications in Computer and Information Science, vol 556, Springer Cham, 2015, p. 16–29. doi:10.1007/978-3-319-25249-0_2
    [BibTeX] [Abstract] [Download PDF]
    In this paper, markerless method of human motion tracking for measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdlers at various training levels. Acquisition of video sequences was carried out under simulated starting conditions of a 110 m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth poses. As the quality criterion, the mean absolute error was adopted. The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.
    @incollection{Krzeszowski2015,
    abstract = {In this paper, markerless method of human motion tracking for measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdlers at various training levels. Acquisition of video sequences was carried out under simulated starting conditions of a 110 m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth poses. As the quality criterion, the mean absolute error was adopted. The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.},
    author = {Krzeszowski, Tomasz and Przednowek, Krzysztof and Iskra, Janusz and Wiktorowicz, Krzysztof},
    booktitle = {Sports Science Research and Technology Support. icSPORTS 2014. Communications in Computer and Information Science, vol 556},
    doi = {10.1007/978-3-319-25249-0_2},
    isbn = {978-3-319-25248-3},
    pages = {16--29},
    publisher = {Springer Cham},
    title = {{Monocular Tracking of Human Motion in Evaluation of Hurdle Clearance}},
    url = {http://dx.doi.org/10.1007/978-3-319-25249-0_2},
    year = {2015}
    }
  • K. Wiktorowicz, K. Przednowek, L. Lassota, and T. Krzeszowski, „Predictive modeling in race walking,” Computational Intelligence and Neuroscience, vol. 2015, 2015. doi:10.1155/2015/735060
    [BibTeX] [Abstract] [Download PDF]
    This paper presents the use of linear and nonlinear multivariable models as tools to support training process of race walkers. These models are calculated using data collected from race walkers' training events and they are used to predict the result over a 3 km race based on training loads. The material consists of 122 training plans for 21 athletes. In order to choose the best model leave-one-out cross-validation method is used. The main contribution of the paper is to propose the nonlinear modifications for linear models in order to achieve smaller prediction error. It is shown that the best model is a modified LASSO regression with quadratic terms in the nonlinear part. This model has the smallest prediction error and simplified structure by eliminating some of the predictors.
    @article{Wiktorowicz2015,
    abstract = {This paper presents the use of linear and nonlinear multivariable models as tools to support training process of race walkers. These models are calculated using data collected from race walkers' training events and they are used to predict the result over a 3 km race based on training loads. The material consists of 122 training plans for 21 athletes. In order to choose the best model leave-one-out cross-validation method is used. The main contribution of the paper is to propose the nonlinear modifications for linear models in order to achieve smaller prediction error. It is shown that the best model is a modified LASSO regression with quadratic terms in the nonlinear part. This model has the smallest prediction error and simplified structure by eliminating some of the predictors.},
    author = {Wiktorowicz, Krzysztof and Przednowek, Krzysztof and Lassota, Leslaw and Krzeszowski, Tomasz},
    doi = {10.1155/2015/735060},
    journal = {Computational Intelligence and Neuroscience},
    title = {{Predictive modeling in race walking}},
    url = {http://dx.doi.org/10.1155/2015/735060},
    volume = {2015},
    year = {2015}
    }
  • T. Krzeszowski, K. Przednowek, J. Iskra, and K. Wiktorowicz, „Monocular Tracking of Human Motion in Evaluation of Hurdle Clearance,” in Communications in Computer and Information Science, J. Cabri, J. Barreiros, and P. P. Correia, Eds., Springer, 2015, vol. 556, p. 16–29. doi:10.1007/978-3-319-25249-0_2
    [BibTeX] [Abstract] [Download PDF]
    {\textcopyright} Springer International Publishing Switzerland 2015. In this paper, markerless method of human motion tracking for measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdlers at various training levels. Acquisition of video sequences was carried out under simulated starting conditions of a 110m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth poses. As the quality criterion, the mean absolute error was adopted. The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.
    @incollection{Krzeszowski2015a,
    abstract = {{\textcopyright} Springer International Publishing Switzerland 2015. In this paper, markerless method of human motion tracking for measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdlers at various training levels. Acquisition of video sequences was carried out under simulated starting conditions of a 110m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth poses. As the quality criterion, the mean absolute error was adopted. The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.},
    author = {Krzeszowski, Tomasz and Przednowek, Krzysztof and Iskra, Janusz and Wiktorowicz, Krzysztof},
    booktitle = {Communications in Computer and Information Science},
    doi = {10.1007/978-3-319-25249-0_2},
    editor = {Cabri, Jan and Barreiros, Jo{\~{a}}o and Correia, Pedro Pezarat},
    isbn = {9783319252483},
    issn = {1865-0929},
    keywords = {Hurdle clearance,Markerless human motion tracking,Monocular camera,Particle swarm optimization},
    pages = {16--29},
    publisher = {Springer},
    title = {{Monocular Tracking of Human Motion in Evaluation of Hurdle Clearance}},
    url = {http://link.springer.com/10.1007/978-3-319-25249-0_2},
    volume = {556},
    year = {2015}
    }
  • P. Lenik, T. Krzeszowski, K. Przednowek, and J. Lenik, „The Analysis of Basketball Free Throw Trajectory using PSO Algorithm,” in Proceedings of the 3rd International Congress on Sport Sciences Research and Technology Support (icSPORTS 2015), 2015, p. 250–256. doi:10.5220/0005611002500256
    [BibTeX] [Abstract] [Download PDF]
    The following paper described the method for automatic measurement of selected parameters of a basketball free throw trajectory. The research material was based on 10 sequences recorded by a monocular camera. For tracking the ball the particle swarm optimization (PSO) algorithm was used. Additionally the method of ball detection was developed. The study was conducted on a group of 10 basketball players who participated in the Polish Second Division during the 2014/2015 season. The 10 parameters (four distances, three velocities, and three angle parameters) were taken into account. The experimental results showed that the value of the initial angle was equal to 47.27±4.42 degrees, and the height of ball trajectory was at the level of 3.84±0.34 m. The correlation between body height and parameter of a free throw was also determined. The analysis conducted showed a significant correlation between the height and shape of a free throw trajectory. The suggested method can be used in the training process as a tool to improve performance of the free throw.
    @inproceedings{Lenik2015,
    abstract = {The following paper described the method for automatic measurement of selected parameters of a basketball free throw trajectory. The research material was based on 10 sequences recorded by a monocular camera. For tracking the ball the particle swarm optimization (PSO) algorithm was used. Additionally the method of ball detection was developed. The study was conducted on a group of 10 basketball players who participated in the Polish Second Division during the 2014/2015 season. The 10 parameters (four distances, three velocities, and three angle parameters) were taken into account. The experimental results showed that the value of the initial angle was equal to 47.27±4.42 degrees, and the height of ball trajectory was at the level of 3.84±0.34 m. The correlation between body height and parameter of a free throw was also determined. The analysis conducted showed a significant correlation between the height and shape of a free throw trajectory. The suggested method can be used in the training process as a tool to improve performance of the free throw.},
    author = {Lenik, Pawel and Krzeszowski, Tomasz and Przednowek, Krzysztof and Lenik, Justyna},
    booktitle = {Proceedings of the 3rd International Congress on Sport Sciences Research and Technology Support (icSPORTS 2015)},
    doi = {10.5220/0005611002500256},
    isbn = {978-989-758-159-5},
    keywords = {Ball trajectory,Basketball,Object tracking,Particle swarm optimization},
    pages = {250--256},
    publisher = {Scitepress},
    title = {{The Analysis of Basketball Free Throw Trajectory using PSO Algorithm}},
    url = {http://dx.doi.org/10.5220/0005611002500256},
    year = {2015}
    }

2014

  • B. Kwolek, T. Krzeszowski, A. Michalczuk, and H. Josinski, „3D gait recognition using spatio-temporal motion descriptors,” in The 6th Asian Conference on Intelligent Information and Database Systems (ACIIDS 2014), 2014, p. 595–604. doi:10.1007/978-3-319-05458-2_61
    [BibTeX] [Abstract] [Download PDF]
    We present a view independent algorithm for 3D human gait recognition. The identification of the person is achieved using motion data obtained by our markerless 3D motion tracking algorithm. We report its tracking accuracy using ground-truth data obtained by a marker-based motion capture system. The classification is done using SVM built on the proposed spatio-temporal motion descriptors. The identification performance was determined using 230 gait cycles performed by 22 persons. The correctly classified ratio achieved by SVM is equal to 93.5{\%} for rank 1 and 99.6{\%} for rank 3. We show that the recognition performance obtained with the spatio-temporal gait signatures is better in comparison to accuracy obtained with tensorial gait data and reduced by MPCA.
    @inproceedings{Kwolek2014,
    abstract = {We present a view independent algorithm for 3D human gait recognition. The identification of the person is achieved using motion data obtained by our markerless 3D motion tracking algorithm. We report its tracking accuracy using ground-truth data obtained by a marker-based motion capture system. The classification is done using SVM built on the proposed spatio-temporal motion descriptors. The identification performance was determined using 230 gait cycles performed by 22 persons. The correctly classified ratio achieved by SVM is equal to 93.5{\%} for rank 1 and 99.6{\%} for rank 3. We show that the recognition performance obtained with the spatio-temporal gait signatures is better in comparison to accuracy obtained with tensorial gait data and reduced by MPCA.},
    author = {Kwolek, Bogdan and Krzeszowski, Tomasz and Michalczuk, Agnieszka and Josinski, Henryk},
    booktitle = {The 6th Asian Conference on Intelligent Information and Database Systems (ACIIDS 2014)},
    doi = {10.1007/978-3-319-05458-2_61},
    isbn = {978-3-319-05457-5},
    pages = {595--604},
    publisher = {Springer Cham},
    title = {{3D gait recognition using spatio-temporal motion descriptors}},
    url = {http://dx.doi.org/10.1007/978-3-319-05458-2_61},
    volume = {8398 LNCS},
    year = {2014}
    }
  • K. Przednowek, T. Krzeszowski, J. Iskra, and K. Wiktorowicz, „Markerless Motion Tracking in Evaluation of Hurdle Clearance Parameters,” in Proceedings of the 2nd International Congress on Sports Sciences Research and Technology Support (icSPORTS-2014), 2014, p. 129–136. doi:10.5220/0005080601290136
    [BibTeX] [Abstract] [Download PDF]
    In this study, implementation of markerless method of human body motion tracking as a tool of measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdle runners at various training levels. Recording of video sequences was carried out under simulated starting conditions of a 110 m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The suggested method does not involve using any special clothes, markers or estimation support techniques. In the study, the basic numerical characteristics of twenty estimated parameters were presented. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth pose. As the quality criterion, the MAE (Mean Absolute Error) was adopted. In the distance parameters, the least error was obtained for the distance between the center of mass (CM) and the hurdle at the first hurdle clearance phase (MAE = 22.0 mm). For the angular parameters, the least error was obtained for the leg angle at the first hurdle clearance phase (MAE = 3.1°). The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.
    @inproceedings{Przednowek2014,
    abstract = {In this study, implementation of markerless method of human body motion tracking as a tool of measurement of hurdle clearance kinematic parameters was presented. The analysis involved 5 hurdle runners at various training levels. Recording of video sequences was carried out under simulated starting conditions of a 110 m hurdle race. Kinematic parameters were determined based on the analysis of images recorded with a 100 Hz monocular camera. The suggested method does not involve using any special clothes, markers or estimation support techniques. In the study, the basic numerical characteristics of twenty estimated parameters were presented. The accuracy of determined hurdle clearance parameters was verified by comparison of estimated poses with the ground truth pose. As the quality criterion, the MAE (Mean Absolute Error) was adopted. In the distance parameters, the least error was obtained for the distance between the center of mass (CM) and the hurdle at the first hurdle clearance phase (MAE = 22.0 mm). For the angular parameters, the least error was obtained for the leg angle at the first hurdle clearance phase (MAE = 3.1°). The level of computed errors showed that the presented method can be used for estimation of hurdle clearance kinematic parameters.},
    author = {Przednowek, Krzysztof and Krzeszowski, Tomasz and Iskra, Janusz and Wiktorowicz, Krzysztof},
    booktitle = {Proceedings of the 2nd International Congress on Sports Sciences Research and Technology Support (icSPORTS-2014)},
    doi = {10.5220/0005080601290136},
    isbn = {978-989-758-057-4},
    keywords = {Computer Vision,Hurdle Clearance,Markerless Motion Tracking,Particle Swarm Optimization},
    pages = {129--136},
    publisher = {Scitepress},
    title = {{Markerless Motion Tracking in Evaluation of Hurdle Clearance Parameters}},
    url = {http://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0005080601290136},
    year = {2014}
    }
  • T. Krzeszowski, A. Switonski, B. Kwolek, H. Josinski, and K. Wojciechowski, „DTW-based gait recognition from recovered 3-D joint angles and inter-ankle distance,” in Int. Conf. on Computer Vision and Graphics 2014 (ICCVG 2014), LNCS, 2014, p. 356–363. doi:10.1007/978-3-319-11331-9_43
    [BibTeX] [Abstract] [Download PDF]
    We present a view independent approach for 3D human gait recognition. The identification of the person is done on the basis of motion estimated by our marker-less 3D motion tracking algorithm. We show tracking performance using ground-truth data acquired by Vicon motion capture system. The identification is achieved by dynamic time warping using both joint angles and inter-joint distances. We show how to calculate approximate Euclidean distance metric between two sets of Euler angles. We compare the correctly classified ratio obtained by DTW built on unit quaternion distance metric and such an Euler angle distance metric. We then show that combining the rotation distances with inter-ankle distances and other person attributes like height leads to considerably better correctly classified ratio.
    @inproceedings{Krzeszowski2014,
    abstract = {We present a view independent approach for 3D human gait recognition. The identification of the person is done on the basis of motion estimated by our marker-less 3D motion tracking algorithm. We show tracking performance using ground-truth data acquired by Vicon motion capture system. The identification is achieved by dynamic time warping using both joint angles and inter-joint distances. We show how to calculate approximate Euclidean distance metric between two sets of Euler angles. We compare the correctly classified ratio obtained by DTW built on unit quaternion distance metric and such an Euler angle distance metric. We then show that combining the rotation distances with inter-ankle distances and other person attributes like height leads to considerably better correctly classified ratio.},
    author = {Krzeszowski, Tomasz and Switonski, Adam and Kwolek, Bogdan and Josinski, Henryk and Wojciechowski, Konrad},
    booktitle = {Int. Conf. on Computer Vision and Graphics 2014 (ICCVG 2014), LNCS},
    doi = {10.1007/978-3-319-11331-9_43},
    isbn = {978-3-319-11330-2},
    pages = {356--363},
    publisher = {Springer Cham},
    title = {{DTW-based gait recognition from recovered 3-D joint angles and inter-ankle distance}},
    url = {http://dx.doi.org/10.1007/978-3-319-11331-9_43},
    volume = {8671},
    year = {2014}
    }
  • K. Przednowek, J. Iskra, and T. Krzeszowski, „The analysis of hurdling steps using an algorithm of computer vision: the case of a well-trained athlete,” Polish J Sport Med, vol. 30, iss. 4, p. 307–313, 2014.
    [BibTeX] [Abstract] [Download PDF]
    Background. In this paper, a markerless method for tracking the motion of a person as a tool to estimate kinematic parameters of a hurdling step was used. Material and methods. The study is conducted on a hurdler of top sports level. Video sequence recording was conducted in simulated conditions of the start in 110 m hurdle race. Kinematic parameters were determined on the basis of the analysis of image sequences recorded using a single camera with the frequency of 100 Hz. The method proposed in this paper does not require any custom-designed clothes, markers or other technologies facilitating estimation. Results. In this study, we present quantitative data on 21 kinematic parameters and their variability throughout the hurdling step. The results confirmed a high level of sports form of the athlete subjected to the study. Conclusions. The proposed method used to estimate kinematic parameters may be used in the assessment of progress in the training of hurdlers, notably in terms of technique preparation.
    @article{Przednowek2014c,
    abstract = {Background. In this paper, a markerless method for tracking the motion of a person as a tool to estimate kinematic parameters of a hurdling step was used. Material and methods. The study is conducted on a hurdler of top sports level. Video sequence recording was conducted in simulated conditions of the start in 110 m hurdle race. Kinematic parameters were determined on the basis of the analysis of image sequences recorded using a single camera with the frequency of 100 Hz. The method proposed in this paper does not require any custom-designed clothes, markers or other technologies facilitating estimation. Results. In this study, we present quantitative data on 21 kinematic parameters and their variability throughout the hurdling step. The results confirmed a high level of sports form of the athlete subjected to the study. Conclusions. The proposed method used to estimate kinematic parameters may be used in the assessment of progress in the training of hurdlers, notably in terms of technique preparation.},
    author = {Przednowek, Krzysztof and Iskra, Janusz and Krzeszowski, Tomasz},
    journal = {Polish J Sport Med},
    number = {4},
    pages = {307--313},
    title = {{The analysis of hurdling steps using an algorithm of computer vision: the case of a well-trained athlete}},
    url = {https://medycynasportowa.edu.pl/resources/html/article/details?id=20073},
    volume = {30},
    year = {2014}
    }

2013

  • T. Krzeszowski, „Human motion tracking using multiple cameras (in Polish),” PhD Thesis, Silesian University of Technology, Gliwice, Poland, 2013.
    [BibTeX]
    @phdthesis{KrzeszowskiPHDThesis2013,
    author = {Krzeszowski, Tomasz},
    school = {Silesian University of Technology, Gliwice, Poland},
    title = {Human motion tracking using multiple cameras (in {P}olish)},
    year = {2013}
    }
  • B. Rymut, B. Kwolek, and T. Krzeszowski, „GPU-accelerated human motion tracking using particle filter combined with PSO,” in Advanced Concepts for Intelligent Vision Systems. ACIVS 2013, 2013, p. 426–437. doi:10.1007/978-3-319-02895-8_38
    [BibTeX] [Abstract] [Download PDF]
    This paper discusses how to combine particle filter (PF) with particle swarm optimization (PSO) to achieve better object tracking. Owing to multi-swarm based mode seeking the algorithm is capable of maintaining multimodal probability distributions and the tracking accuracy is far better than accuracy of PF or PSO. We propose parallel resampling scheme for particle filtering running on GPU. We show the efficiency of the parallel PF-PSO algorithm on 3D model based human motion tracking. The 3D model is rasterized in parallel and single thread processes one column of the image. Such level of parallelism allows us to efficiently utilize the GPU resources and to perform tracking of the full human body at rates of 15 frames per second. The GPU achieves an average speedup of 7.5 over the CPU. For marker-less motion capture system consisting of four calibrated cameras, the computations were conducted on four CPU cores and four GTX GPUs on two cards.
    @inproceedings{Rymut2013,
    abstract = {This paper discusses how to combine particle filter (PF) with particle swarm optimization (PSO) to achieve better object tracking. Owing to multi-swarm based mode seeking the algorithm is capable of maintaining multimodal probability distributions and the tracking accuracy is far better than accuracy of PF or PSO. We propose parallel resampling scheme for particle filtering running on GPU. We show the efficiency of the parallel PF-PSO algorithm on 3D model based human motion tracking. The 3D model is rasterized in parallel and single thread processes one column of the image. Such level of parallelism allows us to efficiently utilize the GPU resources and to perform tracking of the full human body at rates of 15 frames per second. The GPU achieves an average speedup of 7.5 over the CPU. For marker-less motion capture system consisting of four calibrated cameras, the computations were conducted on four CPU cores and four GTX GPUs on two cards.},
    author = {Rymut, Boguslaw and Kwolek, Bogdan and Krzeszowski, Tomasz},
    booktitle = {Advanced Concepts for Intelligent Vision Systems. ACIVS 2013},
    doi = {10.1007/978-3-319-02895-8_38},
    isbn = {978-3-319-02894-1},
    pages = {426--437},
    publisher = {Springer Cham},
    title = {{GPU-accelerated human motion tracking using particle filter combined with PSO}},
    url = {http://dx.doi.org/10.1007/978-3-319-02895-8_38},
    volume = {8192 LNCS},
    year = {2013}
    }
  • T. Krzeszowski, A. Michalczuk, B. Kwolek, A. Switonski, and H. Josinski, „Gait recognition based on marker-less 3D motion capture,” in 10th IEEE Int. Conf. on Advanced Video and Signal Based Surveillance 2013 (AVSS 2013), 2013, p. 232–237. doi:10.1109/AVSS.2013.6636645
    [BibTeX] [Abstract] [Download PDF]
    We present an algorithm for view-independent gait-based person identification. The identification is achieved using data obtained by our marker-less 3D motion tracking algorithm. The motion tracking was accomplished by a particle swarm optimization algorithm. The accuracy of the motion tracking algorithm was evaluated using ground-truth data from MoCap. It was determined on 88 sequences with 22 walking performers. We obtained 90{\%} identification accuracy (rank 1) on 230 gait cycles.
    @inproceedings{Krzeszowski2013,
    abstract = {We present an algorithm for view-independent gait-based person identification. The identification is achieved using data obtained by our marker-less 3D motion tracking algorithm. The motion tracking was accomplished by a particle swarm optimization algorithm. The accuracy of the motion tracking algorithm was evaluated using ground-truth data from MoCap. It was determined on 88 sequences with 22 walking performers. We obtained 90{\%} identification accuracy (rank 1) on 230 gait cycles.},
    author = {Krzeszowski, Tomasz and Michalczuk, Agnieszka and Kwolek, Bogdan and Switonski, Adam and Josinski, Henryk},
    booktitle = {10th IEEE Int. Conf. on Advanced Video and Signal Based Surveillance 2013 (AVSS 2013)},
    doi = {10.1109/AVSS.2013.6636645},
    isbn = {978-1-4799-0703-8},
    pages = {232--237},
    publisher = {IEEE},
    title = {{Gait recognition based on marker-less 3D motion capture}},
    url = {http://dx.doi.org/10.1109/AVSS.2013.6636645},
    year = {2013}
    }

2012

  • T. Krzeszowski, B. Kwolek, A. Michalczuk, A. Switonski, and H. Josinski, „View independent human gait recognition using markerless 3d human motion capture,” in Int. Conf. on Computer Vision and Graphics 2012 (ICCVG 2012), 2012, p. 491–500. doi:10.1007/978-3-642-33564-8_59
    [BibTeX] [Abstract] [Download PDF]
    We present an algorithm for view-independent human gait recognition. The human gait recognition is achieved using data obtained by our markerless 3D motion tracking algorithm. The tensorial gait data were reduced by multilinear principal component analysis and subsequently classified. The performance of the motion tracking algorithm was evaluated using ground-truth data from MoCap. The classification accuracy was determined using video sequences with walking performers. Experiments on multiview video sequences show the promising effectiveness of the proposed algorithm.
    @inproceedings{Krzeszowski2012,
    abstract = {We present an algorithm for view-independent human gait recognition. The human gait recognition is achieved using data obtained by our markerless 3D motion tracking algorithm. The tensorial gait data were reduced by multilinear principal component analysis and subsequently classified. The performance of the motion tracking algorithm was evaluated using ground-truth data from MoCap. The classification accuracy was determined using video sequences with walking performers. Experiments on multiview video sequences show the promising effectiveness of the proposed algorithm.},
    author = {Krzeszowski, Tomasz and Kwolek, Bogdan and Michalczuk, Agnieszka and Switonski, Adam and Josinski, Henryk},
    booktitle = {Int. Conf. on Computer Vision and Graphics 2012 (ICCVG 2012)},
    doi = {10.1007/978-3-642-33564-8_59},
    isbn = {978-3-642-33563-1},
    issn = {0302-9743},
    pages = {491--500},
    publisher = {Springer Berlin Heidelberg},
    title = {{View independent human gait recognition using markerless 3d human motion capture}},
    url = {http://dx.doi.org/10.1007/978-3-642-33564-8_59},
    volume = {7594 LNCS},
    year = {2012}
    }
  • B. Kwolek, T. Krzeszowski, A. Gagalowicz, K. Wojciechowski, and H. Josinski, „Real-time multi-view human motion tracking using particle swarm optimization with resampling,” in 7th Int. Conf. on Articulated Motion and Deformable Objects 2012, 2012, p. 92–101. doi:10.1007/978-3-642-31567-1_9
    [BibTeX] [Abstract] [Download PDF]
    In this paper we propose a particle swarm optimization with resampling for marker-less body tracking. The resampling is employed to select a record of the best particles according to the weights of particles making up the swarm. The algorithm better copes with noise and reduces the premature stagnation. Experiments on 4-camera datasets show the robustness and accuracy of our method. It was evaluated on nine sequences using ground truth provided by Vicon. The full body motion tracking was conducted in real-time on two PC nodes, each of them with two multi-core CPUs with hyper-threading, connected by 1 GigE.
    @inproceedings{Kwolek2012,
    abstract = {In this paper we propose a particle swarm optimization with resampling for marker-less body tracking. The resampling is employed to select a record of the best particles according to the weights of particles making up the swarm. The algorithm better copes with noise and reduces the premature stagnation. Experiments on 4-camera datasets show the robustness and accuracy of our method. It was evaluated on nine sequences using ground truth provided by Vicon. The full body motion tracking was conducted in real-time on two PC nodes, each of them with two multi-core CPUs with hyper-threading, connected by 1 GigE.},
    author = {Kwolek, B. and Krzeszowski, T. and Gagalowicz, A. and Wojciechowski, K. and Josinski, H.},
    booktitle = {7th Int. Conf. on Articulated Motion and Deformable Objects 2012},
    doi = {10.1007/978-3-642-31567-1_9},
    isbn = {978-3-642-31566-4},
    pages = {92--101},
    publisher = {Springer Berlin Heidelberg},
    title = {{Real-time multi-view human motion tracking using particle swarm optimization with resampling}},
    url = {http://dx.doi.org/10.1007/978-3-642-31567-1_9},
    volume = {7378 LNCS},
    year = {2012}
    }
  • T. Krzeszowski, B. Kwolek, B. Rymut, K. Wojciechowski, and H. Josinski, „Real-time tracking of full-body motion using parallel particle swarm optimization with a pool of best particles,” in Int. Conf. on Artificial Intelligence and Soft Computing 2012, Swarm and Evolutionary Computation, 2012, p. 102–109. doi:10.1007/978-3-642-29353-5_12
    [BibTeX] [Abstract] [Download PDF]
    In this paper we present a particle swarm optimization (PSO) based approach for marker-less full body motion tracking. The objective function is smoothed in an annealing scheme and then quantized. This allows us to extract a pool of candidate best particles. The algorithm selects a global best from such a pool to force the PSO jump out of stagnation. Experiments on 4-camera datasets demonstrate the robustness and accuracy of our method. The tracking is conducted on 2 PC nodes with multi-core CPUs, connected by 1 GigE. This makes our system capable of accurately recovering full body movements with 14 fps.
    @inproceedings{Krzeszowski2012a,
    abstract = {In this paper we present a particle swarm optimization (PSO) based approach for marker-less full body motion tracking. The objective function is smoothed in an annealing scheme and then quantized. This allows us to extract a pool of candidate best particles. The algorithm selects a global best from such a pool to force the PSO jump out of stagnation. Experiments on 4-camera datasets demonstrate the robustness and accuracy of our method. The tracking is conducted on 2 PC nodes with multi-core CPUs, connected by 1 GigE. This makes our system capable of accurately recovering full body movements with 14 fps.},
    archivePrefix = {arXiv},
    arxivId = {1},
    author = {Krzeszowski, Tomasz and Kwolek, Bogdan and Rymut, Boguslaw and Wojciechowski, Konrad and Josinski, Henryk},
    booktitle = {Int. Conf. on Artificial Intelligence and Soft Computing 2012, Swarm and Evolutionary Computation},
    doi = {10.1007/978-3-642-29353-5_12},
    eprint = {1},
    isbn = {978-3-642-29352-8},
    pages = {102--109},
    pmid = {31767271},
    publisher = {Springer Berlin Heidelberg},
    title = {{Real-time tracking of full-body motion using parallel particle swarm optimization with a pool of best particles}},
    url = {http://dx.doi.org/10.1007/978-3-642-29353-5_12},
    volume = {7269 LNCS},
    year = {2012}
    }
  • B. Rymut, T. Krzeszowski, and B. Kwolek, „Full body motion tracking in monocular images using particle swarm optimization,” in Int. Conf. on Artificial Intelligence and Soft Computing 2012, Artificial Intelligence and Soft Computing, 2012, p. 600–607. doi:10.1007/978-3-642-29347-4_70
    [BibTeX] [Abstract] [Download PDF]
    The estimation of full body pose in monocular images is a very difficult problem. In 3D-model based motion tracking the challenges arise as at least one-third of degrees of freedom of the human pose that needs to be recovered is nearly unobservable in any given monocular image. In this paper, we deal with high dimensionality of the search space through estimating the pose in a hierarchical manner using Particle Swarm Optimization. Our method fits the projected body parts of an articulated model to detected body parts at color images with support of edge distance transform. The algorithm was evaluated quantitatively through the use of the motion capture data as ground truth. {\textcopyright} 2012 Springer-Verlag Berlin Heidelberg.
    @inproceedings{Rymut2012,
    abstract = {The estimation of full body pose in monocular images is a very difficult problem. In 3D-model based motion tracking the challenges arise as at least one-third of degrees of freedom of the human pose that needs to be recovered is nearly unobservable in any given monocular image. In this paper, we deal with high dimensionality of the search space through estimating the pose in a hierarchical manner using Particle Swarm Optimization. Our method fits the projected body parts of an articulated model to detected body parts at color images with support of edge distance transform. The algorithm was evaluated quantitatively through the use of the motion capture data as ground truth. {\textcopyright} 2012 Springer-Verlag Berlin Heidelberg.},
    author = {Rymut, Boguslaw and Krzeszowski, Tomasz and Kwolek, Bogdan},
    booktitle = {Int. Conf. on Artificial Intelligence and Soft Computing 2012, Artificial Intelligence and Soft Computing},
    doi = {10.1007/978-3-642-29347-4_70},
    isbn = {978-3-642-29346-7},
    number = {PART 1},
    pages = {600--607},
    publisher = {Springer Berlin Heidelberg},
    title = {{Full body motion tracking in monocular images using particle swarm optimization}},
    url = {http://dx.doi.org/10.1007/978-3-642-29347-4_70},
    volume = {7267 LNAI},
    year = {2012}
    }

2011

  • T. Krzeszowski, B. Kwolek, and K. Wojciechowski, „Model-Based 3D Human Motion Capture Using Global-Local Particle Swarm Optimizations,” in Int. Conference on Computer Recognition Systems 2011, Computer Recognition Systems 4, Springer Berlin Heidelberg, 2011, vol. 95, p. 297–306. doi:10.1007/978-3-642-20320-6_31
    [BibTeX] [Abstract] [Download PDF]
    We present an approach for tracking the articulated motion of humans using image sequences obtained from multiple calibrated cameras. A 3D human body model composed of eleven segments, which allows both rotation at joints and translation, is utilized to estimate the pose. We assume that the initial pose estimate is available. A modified swarm intelligence based searching scheme is utilized to perform motion tracking. At the beginning of each optimization cycle, we estimate the pose of the whole body and then we refine locally the limb poses using smaller number of particles. The results that were achieved in our experiments are compared with those produced by other state-of-the-art methods, with analyses carried out both through qualitative visual evaluations as well as quantitatively by the use of the motion capture data as ground truth. They indicate that our method outperforms the algorithm based on the ordinary particle swarm optimization.
    @incollection{Krzeszowski2011,
    abstract = {We present an approach for tracking the articulated motion of humans using image sequences obtained from multiple calibrated cameras. A 3D human body model composed of eleven segments, which allows both rotation at joints and translation, is utilized to estimate the pose. We assume that the initial pose estimate is available. A modified swarm intelligence based searching scheme is utilized to perform motion tracking. At the beginning of each optimization cycle, we estimate the pose of the whole body and then we refine locally the limb poses using smaller number of particles. The results that were achieved in our experiments are compared with those produced by other state-of-the-art methods, with analyses carried out both through qualitative visual evaluations as well as quantitatively by the use of the motion capture data as ground truth. They indicate that our method outperforms the algorithm based on the ordinary particle swarm optimization.},
    author = {Krzeszowski, Tomasz and Kwolek, Bogdan and Wojciechowski, Konrad},
    booktitle = {Int. Conference on Computer Recognition Systems 2011, Computer Recognition Systems 4},
    doi = {10.1007/978-3-642-20320-6_31},
    isbn = {978-3-642-20319-0},
    issn = {1867-5662},
    number = {4},
    pages = {297--306},
    publisher = {Springer Berlin Heidelberg},
    title = {{Model-Based 3D Human Motion Capture Using Global-Local Particle Swarm Optimizations}},
    url = {http://dx.doi.org/10.1007/978-3-642-20320-6_31},
    volume = {95},
    year = {2011}
    }
  • B. Kwolek, T. Krzeszowski, and K. Wojciechowski, „Real-time multi-view human motion tracking using 3D model and latency tolerant parallel particle swarm optimization,” in Int. Conf. on Computer Vision/Computer Graphics Collaboration Techniques „MIRAGE 2011”, 2011, p. 169–180. doi:10.1007/978-3-642-24136-9_15
    [BibTeX] [Abstract] [Download PDF]
    This paper demonstrates how latency tolerant parallel particle swarm optimization can be used to achieve real-time full-body motion tracking. The tracking is realized using multi-view images and articulated 3D model with a truncated cones-based representation of the body. Each CPU core computes fitness score for a single camera. On each node the algorithm uses the current temporary best fitness value without waiting for the global best one from cooperating sub-swarms. The algorithm runs at 10 Hz on eight PC nodes connected by 1 GigE.
    @inproceedings{Kwolek2011,
    abstract = {This paper demonstrates how latency tolerant parallel particle swarm optimization can be used to achieve real-time full-body motion tracking. The tracking is realized using multi-view images and articulated 3D model with a truncated cones-based representation of the body. Each CPU core computes fitness score for a single camera. On each node the algorithm uses the current temporary best fitness value without waiting for the global best one from cooperating sub-swarms. The algorithm runs at 10 Hz on eight PC nodes connected by 1 GigE.},
    author = {Kwolek, Bogdan and Krzeszowski, Tomasz and Wojciechowski, Konrad},
    booktitle = {Int. Conf. on Computer Vision/Computer Graphics Collaboration Techniques "MIRAGE 2011"},
    doi = {10.1007/978-3-642-24136-9_15},
    isbn = {978-3-642-24135-2},
    pages = {169--180},
    publisher = {Springer Berlin Heidelberg},
    title = {{Real-time multi-view human motion tracking using 3D model and latency tolerant parallel particle swarm optimization}},
    url = {http://dx.doi.org/10.1007/978-3-642-24136-9_15},
    volume = {6930 LNCS},
    year = {2011}
    }
  • B. Kwolek, T. Krzeszowski, and K. Wojciechowski, „Swarm intelligence based searching schemes for articulated 3D body motion tracking,” in Advanced Concepts for Intelligent Vision Systems 2011, 2011, p. 115–126. doi:10.1007/978-3-642-23687-7_11
    [BibTeX] [Abstract] [Download PDF]
    We investigate swarm intelligence based searching schemes for effective articulated human body tracking. The fitness function is smoothed in an annealing scheme and then quantized. This allows us to extract a pool of candidate best particles. The algorithm selects a global best from such a pool. We propose a global-local annealed particle swarm optimization to alleviate the inconsistencies between the observed human pose and the estimated configuration of the 3D model. At the beginning of each optimization cycle, estimation of the pose of the whole body takes place and then the limb poses are refined locally using smaller number of particles. The investigated searching schemes were compared by analyses carried out both through qualitative visual evaluations as well as quantitatively through the use of the motion capture data as ground truth. The experimental results show that our algorithm outperforms the other swarm intelligence searching schemes. The images were captured using multi-camera system consisting of calibrated and synchronized cameras. {\textcopyright} 2011 Springer-Verlag.
    @inproceedings{Kwolek2011a,
    abstract = {We investigate swarm intelligence based searching schemes for effective articulated human body tracking. The fitness function is smoothed in an annealing scheme and then quantized. This allows us to extract a pool of candidate best particles. The algorithm selects a global best from such a pool. We propose a global-local annealed particle swarm optimization to alleviate the inconsistencies between the observed human pose and the estimated configuration of the 3D model. At the beginning of each optimization cycle, estimation of the pose of the whole body takes place and then the limb poses are refined locally using smaller number of particles. The investigated searching schemes were compared by analyses carried out both through qualitative visual evaluations as well as quantitatively through the use of the motion capture data as ground truth. The experimental results show that our algorithm outperforms the other swarm intelligence searching schemes. The images were captured using multi-camera system consisting of calibrated and synchronized cameras. {\textcopyright} 2011 Springer-Verlag.},
    author = {Kwolek, Bogdan and Krzeszowski, Tomasz and Wojciechowski, Konrad},
    booktitle = {Advanced Concepts for Intelligent Vision Systems 2011},
    doi = {10.1007/978-3-642-23687-7_11},
    isbn = {978-3-642-23686-0},
    pages = {115--126},
    publisher = {Springer Berlin Heidelberg},
    title = {{Swarm intelligence based searching schemes for articulated 3D body motion tracking}},
    url = {http://dx.doi.org/10.1007/978-3-642-23687-7_11},
    volume = {6915 LNCS},
    year = {2011}
    }
  • T. Krzeszowski and B. Kwolek, „An approach for model-based 3D human pose tracking, animation and evaluation,” in Int. Conf. on Image Processing and Communications 2011, Advances in Intelligent and Soft Computing, Springer Berlin Heidelberg, 2011, vol. 102, p. 173–181. doi:10.1007/978-3-642-23154-4_20
    [BibTeX] [Abstract] [Download PDF]
    This work presents an approach for 3D human pose tracking, animation and evaluation. The tracking of the full body is done using a modified particle swarm optimization and two synchronized cameras. On the basis of the 3D pose estimates we generate animated human motion. The animated images are processed in the same way as videos taken from the CCD cameras. This way we obtained ground-truth and utilized it in evaluations of the motion tracker.
    @incollection{Krzeszowski2011a,
    abstract = {This work presents an approach for 3D human pose tracking, animation and evaluation. The tracking of the full body is done using a modified particle swarm optimization and two synchronized cameras. On the basis of the 3D pose estimates we generate animated human motion. The animated images are processed in the same way as videos taken from the CCD cameras. This way we obtained ground-truth and utilized it in evaluations of the motion tracker.},
    author = {Krzeszowski, Tomasz and Kwolek, Bogdan},
    booktitle = {Int. Conf. on Image Processing and Communications 2011, Advances in Intelligent and Soft Computing},
    doi = {10.1007/978-3-642-23154-4_20},
    isbn = {978-3-642-23153-7},
    pages = {173--181},
    publisher = {Springer Berlin Heidelberg},
    title = {{An approach for model-based 3D human pose tracking, animation and evaluation}},
    url = {http://dx.doi.org/10.1007/978-3-642-23154-4_20},
    volume = {102},
    year = {2011}
    }
  • T. Krzeszowski, B. Kwolek, K. Wojciechowski, and H. Josinski, „Markerless articulated human body tracking for gait analysis and recognition,” Machine Graphics & Vision, vol. 20, iss. 3, p. 267–281, 2011.
    [BibTeX] [Abstract] [Download PDF]
    We present a particle swarm optimization (PSO) based system for markerless full body motion tracking. The fitness function is smoothed in an annealing scheme and then quantized. In this manner we extract a pool of candidate best particles. The swarm of particles selects a global best from such a pool of the particles to force the PSO the jump out of stagnation. Experiments on 4-camera datasets demonstrate the accuracy of our method on image sequences with walking persons. The system was evaluated using ground-truth data from a marker-based motion capture system by Vicon. We compared the joint motions and the distances between ankles, which were extracted using both systems. Thanks to the high precision of the markerless motion estimation, the curves illustrating the distances between ankles overlap considerably in almost all frames of the image sequences.
    @article{Krzeszowski2011b,
    abstract = {We present a particle swarm optimization (PSO) based system for markerless full body motion tracking. The fitness function is smoothed in an annealing scheme and then quantized. In this manner we extract a pool of candidate best particles. The swarm of particles selects a global best from such a pool of the particles to force the PSO the jump out of stagnation. Experiments on 4-camera datasets demonstrate the accuracy of our method on image sequences with walking persons. The system was evaluated using ground-truth data from a marker-based motion capture system by Vicon. We compared the joint motions and the distances between ankles, which were extracted using both systems. Thanks to the high precision of the markerless motion estimation, the curves illustrating the distances between ankles overlap considerably in almost all frames of the image sequences.},
    author = {Krzeszowski, T. and Kwolek, B. and Wojciechowski, K. and Josinski, H.},
    journal = {Machine Graphics {\&} Vision},
    keywords = {Gait analysis and recognition,Markerless motion tracking,Particle swarm optimization},
    number = {3},
    pages = {267--281},
    title = {{Markerless articulated human body tracking for gait analysis and recognition}},
    url = {http://mgv.wzim.sggw.pl/MGV20.html},
    volume = {20},
    year = {2011}
    }

2010

  • T. Krzeszowski, B. Kwolek, and K. Wojciechowski, Articulated body motion tracking by combined particle swarm optimization and particle filtering, Springer Berlin Heidelberg, 2010, vol. 6374 LNCS. doi:10.1007/978-3-642-15910-7_17
    [BibTeX] [Abstract] [Download PDF]
    This paper proposes the use of a particle filter with embedded particle swarm optimization as an efficient and effective way of dealing with 3d model-based human body tracking. A particle swarm optimization algorithm is utilized in the particle filter to shift the particles toward more promising configurations of the human model. The algorithm is shown to be able of tracking full articulated body motion efficiently. It outperforms the annealed particle filter, kernel particle filter as well as a tracker based on particle swarm optimization. Experiments on real video sequences as well as a qualitative analysis demonstrate the strength of the approach. {\textcopyright} 2010 Springer-Verlag Berlin Heidelberg.
    @book{Krzeszowski2010,
    abstract = {This paper proposes the use of a particle filter with embedded particle swarm optimization as an efficient and effective way of dealing with 3d model-based human body tracking. A particle swarm optimization algorithm is utilized in the particle filter to shift the particles toward more promising configurations of the human model. The algorithm is shown to be able of tracking full articulated body motion efficiently. It outperforms the annealed particle filter, kernel particle filter as well as a tracker based on particle swarm optimization. Experiments on real video sequences as well as a qualitative analysis demonstrate the strength of the approach. {\textcopyright} 2010 Springer-Verlag Berlin Heidelberg.},
    author = {Krzeszowski, T. and Kwolek, B. and Wojciechowski, K.},
    booktitle = {Int. Conf. on Computer Vision and Graphics 2010},
    doi = {10.1007/978-3-642-15910-7_17},
    isbn = {978-3-642-15909-1},
    number = {PART 1},
    pages = {147--154},
    publisher = {Springer Berlin Heidelberg},
    title = {{Articulated body motion tracking by combined particle swarm optimization and particle filtering}},
    url = {http://dx.doi.org/10.1007/978-3-642-15910-7_17},
    volume = {6374 LNCS},
    year = {2010}
    }
  • T. Krzeszowski, B. Kwolek, and K. Wojciechowski, „GPU-accelerated tracking of the motion of 3D articulated figure,” in Int. Conf. on Computer Vision and Graphics 2010, 2010, p. 155–162. doi:10.1007/978-3-642-15910-7_18
    [BibTeX] [Abstract] [Download PDF]
    This paper presents methods that utilize the advantages of modern graphics card hardware for real-time full body tracking with a 3D body model. By means of the presented methods the tracking of full body can be performed at frame-rates of 5 frames per second using a single low-cost moderately-priced graphics card and images from single camera. For a model with 26 DOF we achieved 15 times speed-up. The pose configuration is given by the position and orientation of the pelvis as well as relative joint angles between the connected limbs. The tracking is done through searching for a model configuration that best corresponds to the observed human silhouette in the input image. The searching is done via particle swarm optimization, where each particle corresponds to some hypothesized set of model parameters. {\textcopyright} 2010 Springer-Verlag Berlin Heidelberg.
    @inproceedings{Krzeszowski2010a,
    abstract = {This paper presents methods that utilize the advantages of modern graphics card hardware for real-time full body tracking with a 3D body model. By means of the presented methods the tracking of full body can be performed at frame-rates of 5 frames per second using a single low-cost moderately-priced graphics card and images from single camera. For a model with 26 DOF we achieved 15 times speed-up. The pose configuration is given by the position and orientation of the pelvis as well as relative joint angles between the connected limbs. The tracking is done through searching for a model configuration that best corresponds to the observed human silhouette in the input image. The searching is done via particle swarm optimization, where each particle corresponds to some hypothesized set of model parameters. {\textcopyright} 2010 Springer-Verlag Berlin Heidelberg.},
    author = {Krzeszowski, T. and Kwolek, B. and Wojciechowski, K.},
    booktitle = {Int. Conf. on Computer Vision and Graphics 2010},
    doi = {10.1007/978-3-642-15910-7_18},
    isbn = {978-3-642-15909-1},
    number = {PART 1},
    pages = {155--162},
    publisher = {Springer Berlin Heidelberg},
    title = {{GPU-accelerated tracking of the motion of 3D articulated figure}},
    url = {http://dx.doi.org/10.1007/978-3-642-15910-7_18},
    volume = {6374 LNCS},
    year = {2010}
    }
Close Menu