Abstract
In recent years, acupuncture and moxibustion techniques of traditional Chinese medicine have been developed rapidly, while acupuncture and moxibustion techniques of traditional Mongolian medicine still have room for further development. How to accurately and quickly find the location of Mongolian acupuncture points is a difficult point in the inheritance and practice of Mongolian acupuncture and moxibustion. This paper proposes a method of Mongolian medicine hand acupoint recognition based on machine vision. First, use the YOLOv3 model to judge whether it is the palm or the back of the hand. Then, use MediaPipe to obtain the coordinates of 21 hand key points, and calculate the distance and angle relationship between Mongolian medicine hand acupoints and hand key points. Finally, the coordinates of the acupoints on the hands of Mongolian medicine are calculated. A comparative experiment was carried out between the data calibrated by Mongolian medical experts and the data calibrated by the recognition method. Experiments show that this method has high recognition rate and good robustness.