MatchMentor uses AI in order to analyse videos of football games. To do this, we use deep convolution neural nets which filter images to find revelant features which are used to determine which team has possession and when certain events occured during the match. This is completed by passing images through the network frame-by-frame, predicting events of interest, calculating the errors for these predictions, and using algorithms to update the network to minimise these errors to gain more accurate results for predictions on subsequent frames. This is then all applied within a graphical user interface to allow users to access the analysis outputs without the need of programming knowledge.

This project democratises sport training in order to level the playing field for all athletes who may not have access to elite or academy level coaching resources, so that players can progress to their full potential no matter their level or financial situations.