The Improved Hu Moment and Its Application in Gesture Recognition
Tianze Zhang, Xinghua Gao, Jianyong Li · 2020
Hu moment is a classic gesture recognition algorithm. Although its calculation is simple, the accuracy is low. In order to solve the problem of low accuracy of traditional Hu moment recognition image features, this paper improves the algorithm by changing the characteristic value of Hu algorithm and calculating the similarity between the image to be recognized and the template image. This paper use Kinect sensors to collect gesture information. After gesture recognition processing, extract hand contour information in order to track important palm nodes. Experimental results show that the improved algorithm's gesture recognition accuracy is higher than 95%, which can effectively improve the gesture recognition accuracy.