A table tennis performance analyzer via a single-view low-quality camera

This paper proposes an approach to analyze the performance of a table tennis player in returning balls launched by a machine. Video frames streamed by a single-view low-quality camera are used as the input of the proposed approach. Background subtraction, HSV transformation, and morphological processing are also applied to identify the location of the ball. The output of the proposed approach is types of balls returned by the player. Seven hundred and fifty balls were launched to 15 different zones on the player's course in the experiment. The result shows that the accuracy of the proposed approach is 96.29%.