A novel realtime vision-based acupoint estimation for TCM massage robot

Abstract

Traditional Chinese medicine(TCM) massage robot is a new technology that combines traditional Chinese medicine and modern robot. It plays a vital role in alleviating the shortage of doctor resources and achieving the clinical curative effect of traditional Chinese medicine. Self-localization of acupoint massage points is the key to realizing TCM massage robots' automatic and intelligent massage. A novel real-time 3D visual acupoint estimation method is proposed in this paper. Firstly, the neural network is used to obtain the information of human joints, then proportional geometry measurement is used to analyze the geometric relationship of acupoints, and finally, the accurate positioning of acupoints is realized combined with the 3D information. This method can realize fully automatic real-time acupoint 3D estimation. The research shows that the average positioning accuracy of this method is 2.36cm, the real-time performance can reach 14fps, and it has a good adaptive effect for acupoint positioning of people with different bodies, environments, and postures.