Feature Extraction of Collegiate Gymnastics Difficulty Movements Based on Radon Transform

In this paper, an innovative method combining Radon transform and time convolutional network (TCN) is proposed to meet the requirement of difficult movement feature extraction in college students’ aerobics. With the popularity of calisthenics among college students, the accurate evaluation of the difficulty of aerobics has become the key to improve the training effect and competitive performance. However, the existing motion analysis methods often ignore the time series characteristics of the motion, which limits the comprehensive understanding of the complexity of the motion. To this end, we propose a novel methodological framework, which first extracts key linear features of calisthenics movements by using Radon transform, and then analyzes the time series features of these movements by using TCN, so as to fully capture the complexity of movements. With this approach, we aren’t solely able to identify the key gestures within the movement, but additionally to evaluate the development procedure of the movement and the price of change. In the experimental part, we gather multi-perspective movies of university students’ calisthenics and apply the proposed characteristic extraction approach to them after preprocessing. The experimental results display that, in comparison with conventional strategies, our approach can pick out and examine the issue level of aerobics extra accurately. Mainly, within the undertaking of motion difficulty class, the method combining Radon remodel and TCN indicates tremendous performance improvement, which verifies its software capacity in the area of motion evaluation. This have a look at not only provides a powerful technical method for the characteristic extraction and issue assessment of aerobics, but additionally provides a brand new angle and approach for the studies inside the subject of sports technological know-how.