Application of Machine Learning in Sport Game Analysis
Oral Presentation
Paper ID : 1945-12THCONG
Authors
Machine Learning and Signal Processing Research Group University of Tehran
Abstract
Artificial intelligence is continuously developing to benefit many different businesses. Machine learning is one of the intelligent methodologies that have shown promising results in classification and prediction. The broadcasting of sport events and professional athlete are two big businesses, which use data mining and machine learning to attract more customers. Furthermore, sports data mining assists coaches and managers in result prediction, player performance assessment and game strategy evaluation.
The objective of the sports video analysis is to extract semantics from the source video and adapt the results to various applications. There has been significant interest in algorithms for automatic event detection in sports broadcasts. Event detection examines the sports video for events like goal-shootings, fouls, scoring a point, etc. These algorithms are applied on some low-level features such as players’ positions, motion trajectories, and ball actions. Such features are extracted from sport videos, using mathematical models. Different segmentation and motion tracking techniques are used, including color based, graphical, probabilistic and usual feature extraction methods.
In this talk I will introduce some of the more relevant applications of the artificial intelligence and machine learning in sport game analysis. First, I will have a short review on machine learning techniques. Second, I will review some of the feature extraction methods, including shot classification, ball and player tracking, trajectory extraction, object detection and segmentation. Finally, I will introduce some of more relevant applications, including event detection in sport videos, camera calibration, tactics and performance analysis, referee assistance, result prediction, real-time action recognition, augmented reality, game designs and playing strategies.
The objective of the sports video analysis is to extract semantics from the source video and adapt the results to various applications. There has been significant interest in algorithms for automatic event detection in sports broadcasts. Event detection examines the sports video for events like goal-shootings, fouls, scoring a point, etc. These algorithms are applied on some low-level features such as players’ positions, motion trajectories, and ball actions. Such features are extracted from sport videos, using mathematical models. Different segmentation and motion tracking techniques are used, including color based, graphical, probabilistic and usual feature extraction methods.
In this talk I will introduce some of the more relevant applications of the artificial intelligence and machine learning in sport game analysis. First, I will have a short review on machine learning techniques. Second, I will review some of the feature extraction methods, including shot classification, ball and player tracking, trajectory extraction, object detection and segmentation. Finally, I will introduce some of more relevant applications, including event detection in sport videos, camera calibration, tactics and performance analysis, referee assistance, result prediction, real-time action recognition, augmented reality, game designs and playing strategies.
Keywords
Subjects