Player detection and ball tracking in soccer game scenes

• Grant number: 20/02538-1
• Support type: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): April 01, 2020
• Effective date (End): March 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal researcher: Marcos de Sales Guerra Tsuzuki
• Grantee: Fernando Ferreira Doria
• Home Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil

Abstract

The work will create a detection and tracking algorithm for players and the ball in football matches. Players will be detected according to the algorithms: 1. Filter based detection applied in labeled region, 2. Color based detection (RGB or HSV). For detection and tracking of the ball will be used the VGG-M neural network. This will allow a complete analysis of matches, thus extracting relevant information, such as the summary of the game. The Python programming language will be used. The candidate has started his research activities since September 2019, and some results are presented in this document. (AU)