A Video-Based Framework for Automatic 3D Localization of Multiple Basketball Players: A Combinatorial Optimization Approach.

Sports complexity must be investigated at competitions; therefore, non-invasive methods are essential. In this context, computer vision, image processing, and machine learning techniques can be useful in designing a non-invasive system for data acquisition that identifies players' positions in official basketball matches. Here, we propose and evaluate a novel video-based framework to perform automatic 3D localization of multiple basketball players. The introduced framework comprises two parts. The first stage is player detection, which aims to identify players' heads at the camera image level. This stage is based on background segmentation and on classification performed by an artificial neural network. The second stage is related to 3D reconstruction of the player positions from the images provided by the different cameras used in the acquisition. This task is tackled by formulating a constrained combinatorial optimization problem that minimizes the re-projection error while maximizing the number of detections in the formulated 3D localization problem.

Measuring handball players trajectories using an automatically trained boosting algorithm.

The aim of the present paper is to propose and evaluate an automatically trained cascaded boosting detector algorithm based on morphological segmentation for tracking handball players. The proposed method was able to detect correctly 84% of players when applied to the second period of that same game used for training and 74% when applied to a different game. Furthermore, the analysis of the automatic training using boosting detector revealed general results such as the training time initially increased with the number of figures used, but as more figures were added, the training time decreased. Automatic morphological segmentation has shown to be a fast and efficient method for selecting image regions for the boosting detector and allowed an improvement in the automatic tracking of handball players.

Tracking of wheelchair rugby players in the 2008 Demolition Derby trial.

The kinematics of male wheelchair rugby players (n = 8) during the final match of the 2008 Demolition Derby were assessed using two cameras (10 Hz) and a tracking method based on image-processing techniques. The players' on-court trajectory, total and accumulated distances covered, distances covered during working and stopped game clock, and average velocity were calculated. The players covered 2294.7 +/- 391.5 m in the first half and 2245.4 +/- 431.5 m in the second half. The average velocity was 14.5% greater in the first (1.22 +/- 0.21 m . s(-1)) than in the second half (1.05 +/- 0.20 m . s(-1)). The reduction in velocity was most marked in players with less functional capacity. The working game clock represented 48% of total game time (32 of 66.8 minutes) and the distance covered by all players during working game clock was 60% of the overall distance covered during the whole game (36,323 m). The results provide new information regarding the dynamics of a wheelchair rugby match and should aid coaches in planning effective training and coaching strategies.

Analysis of the distances covered by first division brazilian soccer players obtained with an automatic tracking method.

Methods based on visual estimation still is the most widely used analysis of the distances that is covered by soccer players during matches, and most description available in the literature were obtained using such an approach. Recently, systems based on computer vision techniques have appeared and the very first results are available for comparisons. The aim of the present study was to analyse the distances covered by Brazilian soccer players and compare the results to the European players', both data measured by automatic tracking system. Four regular Brazilian First Division Championship matches between different teams were filmed. Applying a previously developed automatic tracking system (DVideo, Campinas, Brazil), the results of 55 outline players participated in the whole game (n = 55) are presented. The results of mean distances covered, standard deviations (s) and coefficient of variation (cv) after 90 minutes were 10,012 m, s = 1,024 m and cv = 10.2%, respectively. The results of three-way ANOVA according to playing positions, showed that the distances covered by external defender (10642 ± 663 m), central midfielders (10476 ± 702 m) and external midfielders (10598 ± 890 m) were greater than forwards (9612 ± 772 m) and forwards covered greater distances than central defenders (9029 ± 860 m). The greater distances were covered in standing, walking, or jogging, 5537 ± 263 m, followed by moderate-speed running, 1731 ± 399 m; low speed running, 1615 ± 351 m; high-speed running, 691 ± 190 m and sprinting, 437 ± 171 m. Mean distance covered in the first half was 5,173 m (s = 394 m, cv = 7.6%) highly significant greater (p < 0.001) than the mean value 4,808 m (s = 375 m, cv = 7.8%) in the second half. A minute-by-minute analysis revealed that after eight minutes of the second half, player performance has already decreased and this reduction is maintained throughout the second half. Key pointsA novel automatic tracking method was presented. No previous work was found in the literature reporting data of simultaneous trajectories of all soccer players obtained by an automatic tracking method.The study reveals 7% reduction in mean distance covered in the second half and moreover after eight minutes of the second half, player performance has already decreased and this reduction is maintained throughout the second half.