Video Mining in Basketball Shot and Game Analysis

Abstract

The aim of this study is to analyze the footage of basketball games presented to viewers on television. It includes a wide range of activities from identifying players, determining their position, recognizing the ball, hoops, as well as analyzing the shots and determining shot efficacy. The player detection is based on mixture of non-oriented pictorial structures. The detection of body parts is performed by the Support Vector Machines (SVM) algorithm. This paper contains algorithms for detecting player positions of the court, ball position detection and determination of shot. It is achieved by detecting court position and applying spatial transformation. It also includes detection of shot, detection weather shot was successful and position from which shot was taken. All algorithms are tested in large number of frames from different basketball games.