Re-enacting football matches in VR using virtual agents' realistic behaviours.

Abstract

Football matches are often analysed from 2D video footage that lacks any spatial details of the players on the pitch. On the other hand, the real-time positions of the players derived from recorded tracking data are not easily understandable without an immersive and effective visualisation of the players located on a pitch with the ability of the system to review actions over a timeline. To address some of these drawbacks, we propose a novel “Believable Agent Behaviour” (BAB) engine, based on virtual reality equipped with artificial intelligence for the generation of realistic virtual agent animations to enact the data-driven events from football matches. This system implements an immersive, interactive, believable, and automated end-to-end simulation to analyse and replay a football match using immersive 3D graphics.