Research on the Recognition Method of Fuzzy Features in Tai Chi Force Actions

This article addresses the common problem of image blurring occurring during Tai Chi force action image acquisition. It is intended for improvement of the accuracy with which machine systems can extract and identify action features from blurred images. One way to derive the recognition model that unites conventional machine learning with deep learning in Tai Chi action recognition is preprocessing of the existing Tai Chi video frames-denoising, image enhancement, and the partitioning of blurred areas. Mathematical modeling and quantification are performed on fuzzy images of the blurred features. The results of the experiment show that the proposed method substantially increases the accuracy and identification of the recognition, the research oriented around digital analysis, training guidance, and correction of martial arts movements and is an important foundation for other studies in similar areas.