Application of Image Segmenting Technology Based on Fuzzy C-Means Algorithm in Competition Video Referee

Abstract

As sports developing and popularizing, people have put forward increasing demands for fairness and transparency in competitions. However, traditional competition referee systems rely too heavily on human subjective judgment, which often leads to significant subjectivity and in-accuracy issues. To overcome these difficulties, some improvements have been made on the basis of the fuzzy C-means algorithm to preserve more information, while also suppressing membership to improve the convergence speed of the algorithm. These results confirm that the accuracy of the model converged approximately in the 600th round, with an accuracy of 96.1%. Compared with the standard fuzzy C-means algorithm, the convergence speed has been improved by 33% and the accuracy has been improved by 2.1%. This experiment verifies the effectiveness and practicality of the video referee in this competition. This provides new ideas for the development of image segmenting technology and has certain application potential in sports competition refereeing, judicial notarization refereeing, and other fields.

Keywords

• Fuzzy theory
• C-means algorithm
• image segmenting technology
• video image recognition
• video referee