Despite the large amount of studies conducted in the field of human pose estimation and tracking in sports, currently there is a lack of a system which is capable to track player arm movements in real time. Such a system can assist the players to master the right techniques with guaranteed optimal performance of the player. In this paper we propose a robust algorithm to model player arm movements in outdoor sporting activities. The system uses trained cascade object classifier to predict a region of interest for arm in the monocular input video sequence. Optical flow algorithm is employed to extract the motion in that region. Arm region in resultant binary image is later classified using Active Shape Model. The algorithm is tested and validated using several experiments for tracking ball delivery process in cricket as well as for tracking service in volleyball and tennis. The algorithm is capable of classifying and tracking player arm movements with more than 80 percent accuracy, irrespective of the position and background complexities that the real gaming conditions offer.
© Copyright 2015 International Journal of Computer Science in Sport. Sciendo. Kaikki oikeudet pidätetään. / Kääntänyt https://www.DeepL.com/Translator