Video analytics is gaining importance in the sporting world. Soccer clubs around the globe have started to incorporate video analysis methods that other sports, such as basketball and American football, have been using for some time. Currently, soccer video analysis is done manually either by the team itself or by companies with dedicated video analysts. With soccer being the world’s most popular sport, there exists an unmet need to automate and improve analysis of soccer video.
The objective of our project is to develop a system for real-time automated analysis of soccer videos using novel computer vision and machine learning techniques. The analysis will include the detection and recognition of important soccer events and activities in soccer videos. The analysis will ultimately discover team strategies over multiple games and statistical relationship of strategies between teams. The outcome this project will potentially affect many parties including soccer teams, coaches, players, sports broadcast industry, and even end users or fans. A short example of player and ball tracking is shown in the video below.