A multi-view automatic table tennis umpiring framework

This paper presents a low-cost, portable, multi-view table tennis umpiring framework, as viable alternative to the current expensive systems which are almost exclusively restricted elite professional sports. Table has been selected sport evaluate this framework primarily because it comprises many different complex match elements, including service, return and rally governed by strict set of regulations need be umpired. The aim is develop novel methods analyse judge legality such key facets, with ball detection tracking in video frames being integral reliably accurately determining ball’s position flight path during rallies. While low-cost option attractive offers several benefits, technically challenging problem due small number generally low-resolution cameras that used. A camera setup multi-agent system (MAS) presented, computationally lightweight agents detect track ball, create 3D representation predict trajectory, identify facets rally. MAS correctly detects all state transitions seven test sequences minimal latency while processing rate standard computer may unable long rallies real time, potential running on parallel architecture propitious alternative. also scalable, enabling additional pairs deployed achieve enhanced accuracy coverage. affords reform way matches umpired, especially for amateur players, providing an economic objective manner dispute resolution, facility extendible other relevant ball-based analysis mechanism can exploited valuable training tool skill development.