Techniques and applications for soccer video analysis: A survey

Nowadays, soccer is the most popular sport in our society, followed by millions of people. Consequently, many video analysis applications have been developed in the last years to provide information that can be useful for viewers, referees, coaches and players. Some of these applications are focused on specific tasks, such as detecting players, segmenting the field of play, or registering the broadcast images. On the other hand, there are applications aimed at performing tasks of a higher level, such as event detection or game analysis. Here, the most meaningful techniques and applications that have been proposed throughout the last two decades to analyze soccer video sequences are surveyed. The aim of the paper is not to compare the existing techniques, but to represent a comprehensive and organized showcase for the state-of-the-art in the field: as such, it provides a thorough review of the existing types of soccer analysis applications and the techniques used in each one of them, along with the apparent recent technical trends identified from the most recent works, and discuses the challenges in soccer analysis that still remain unsolved.