SIGA Project aims at building an innovative intelligent soccer game analysis and simulation system. An autonomous vision system based on six cameras installed on a real soccer stadium, with 3D player’s detection and tracking capabilities, will be the basis for the system information gathering. The information gathered, regarding ball and players positions and velocities (together with several other game parameters) will be analyzed by using an intelligent autonomous agent. This agent will be able to use artificial intelligence techniques in order to rectify and correct information coming from the image processing and analysis system. The agent will also be capable of calculating complex individual and team performance statistics.

Information gathered during a soccer match includes for individual players: distance covered and running speed, motor tasks (like runs, jumps, receptions, crosses, passes, ball contacts, throw ins, shots, headings) and other information regarding player individual actions. Team performance information includes: channels of ball circulation, ball possession, attacking and defense elapsing time, ball lost, styles and methods of play, zones of player’s interaction, space of effective play, game pace and other complex information regarding team play.

The information gathered by the game analysis agent will be stored in an electronic format appropriate for its analysis and real-time visualization using a graphical system. This graphical system will be a complete soccer analysis application capable of viewing simultaneously the game and its analysis and to deliver several high-level individual and team performance measures to the system user. This will be a very valuable tool for any sport science researcher, enabling to gather, without any interference with the athletes, quantitative performance measures of all individual players and both teams.

The system most innovative characteristic will be its ability for, based on the information gathered concerning each player and team, create realistic player and team physical and behavior models. Having as its basis these models, the system will be able to simulate real soccer games using teams composed by the modeled players, enabling researchers in the sports area to test different soccer tactical organizations and define methodologies to relate the team performance with the playing style.

Parteners: LIACC-FEUP / FCDEF-UP / ISR-Porto / Univ.Aveiro

Interested in this Project? Do you want to do a PhD or MSc thesis in Portugal on a related subject? Please contact me?