Video-Based Rope Skipping Repetition Counting with ResNet Model

Abstract

Video Repetition Counting is one of the important research areas in computer vision. It focuses on estimating the number of repeating actions. In this paper, we propose a method for video-based rope skipping repetition counting that combines the ResNet Model and a counting algorithm. Each frame in the given video is first classified into two categories: upward and downward, describing its current motion status. Then the classification sequence of the video is processed by a statistical counting algorithm to obtain the final repetition number. The experiments on real-world videos show the efficiency of our model.