Efficient Optical Character Recognition System for Automatic Soccer Player's Identification

In this paper, we present a new system of automatic character recognition in order to quickly and reliably identify soccer on the basis of their team number for sport’s scene understanding. First, the proposed approach extracts the number on the player’s jersey by using a chromaticity-based segmentation method. Then, the extracted character recognition is performed by template matching. Hence, the innovative combination of these two techniques leads to a more computationally efficient system for the player identification purpose than the state-of-the-art ones as demonstrated on real-world image large datasets.