Investigation of Rehabilitation Training Research Based on Deep Learning Models

Abstract

The paper will survey the current status of the Deep Learning (DL) technology application and the trend of the rehabilitation training development. Rehabilitation treatment aims to help patients regain their physical function and improve their quality of life. The introduction of artificial intelligence, especially deep learning, provides new technical support for it. This paper focuses on the typical applications of Convolutional Neural Network (CNN) and its hybrid model in rehabilitation training. It covers the body movement identification, rehabilitation assessment, such as virtual reality interactive multiple dimensions. In the discussion section, current challenges are pointed out, such as poor model interpretability, lack of generality and adaptability, and excessive model size. In order to solve the above problems, the introduction of domain experts to guide optimization strategies, such as distillation using domain adaptive methods is proposed and the knowledge. Think deep learning application prospect in the rehabilitation training, but still needs a multidisciplinary collaborative promote its clinical application and promotion.