Seminarios internacionales, días 25 y 26 de abril

La próxima semana recibimos la visita del  profesor Tomasz Krzeszowski de la  Rzeszow University of Technology. Durante su visita, el  profesor va a impartir dos seminarios  en la sala de juntas del DSIC:

* Miércoles 25 de abril a las 15:00 «Computer Vision: Algorithms and Applications»

* Jueves 26 de abril a las 10:00 «Tracking and analysis of human motion»

Breve resumen de su currículum y resumen de los seminarios.

Tomasz Krzeszowski received the MSc degree in computer science from the Rzeszow University of Technology in 2009 and PhD degree in Computer Science from the Silesian University of Technology in 2013. He is currently an assistant professor in the Department of Computer and Control Engineering in the Rzeszow University of Technology. His research interests are in computer vision, human motion tracking, human body pose estimation, parallel processing, and particle swarm optimization algorithms. He has published more than thirty scientific papers, including articles in journals with impact factor.

Tracking and analysis of human motion

Tomasz Krzeszowski

Department of Computer and Control Engineering, Rzeszow University of Technology

Human motion tracking is a very important problem and has many potential applications. The results of studies on this issue may be used, inter alia in medicine, security and surveillance systems, recognition and analysis of activities, production of films and games, as well as in user-friendly interfaces and artificial/augmented reality systems. Currently, commercial solutions commonly used motion tracking systems, which use markers in the process of tracking. However, these systems have many disadvantages, and the costs of their purchase and operation are very high. These potential applications and disadvantages of marker-based motion tracking systems have contributed to the rapid development of markerless human motion tracking systems.

During the presentation, the markerless human motion capture system will be discussed. This system uses a particle swarm optimization (PSO) algorithm, 3D model and image processing methods to estimate the pose of a human body. The modifications of PSO algorithm, that have improved the quality of tracking, and obtained results will also be presented. In the further part of the presentation, research on the application of markerless tracking methods to gait recognition and analysis of athletes’ techniques will be discussed. In the case of gait recognition, three different approaches will be discussed. These will be methods that utilize Multilinear Principal Components Analysis (MPCA), Dynamic Time Warping (DTW), and Spatio-Temporal Motion Descriptors. The results obtained by mentioned methods on a dataset containing 22 characters will be presented. As an example of the application of the markerless tracking system in analysis of athletes’ techniques the method of estimating hurdle clearance parameters will be presented.

Computer Vision: Algorithms and Applications
Tomasz Krzeszowski

Department of Computer and Control Engineering, Rzeszow University of Technology

Computer vision is an interdisciplinary field that deals with how computers can be made for gaining high-level understanding from digital images or videos. The goal of Computer vision is to emulate human vision based on methods of digital image processing. There are three main processing components: image acquisition, image processing and image analysis and understanding. Computer vision methods have many applications, e.g. motion recognition, augmented reality and autonomous cars.

During the presentation, basic issues related to computer vision will be discussed. Main attention will be focused on methods related to image analysis and understanding. The methods of image segmentation, detection of features and description of objects will be discussed. Part of the attention will also be devoted to machine learning issues, as well as description of WEKA software, which contain a collection of machine learning algorithms for data mining tasks. An example of applying image analysis methods to gait recognition will also be presented.