Player Detection in Football Match Videos

This task could be divided into three parts:

  1. Player Detection
  2. Player Classification
  3. Player Tracking

To conduct this firstly, the field and the stands were isolated from each frame to leave only the players. The players were then detected using a OpenCV and MobileNetv2 ensemble model. The detected players were then classified based on the color percentages in the bounding boxes.

Vamsi Nallapareddy
Vamsi Nallapareddy
Research Assistant at University College London

Research Interests: Computational Biology, Bioinformatics, and Deep Learning

Related