Chinese Sign Language Key Action Recognition Based on Extenics Immune Neural Network
Yue Sun, Tiantian Yuan, Junfen Chen, Rui Feng · 2020
In order to solve the problem of low accuracy of key action capture and feature recognition in Chinese sign language human-computer interaction, a large scale continuous sign language data set is established with Kinect equipment. By combining extension analysis theory with immune neural network (INN), extension immune detector is designed to locate the key actions of the sign language effectively. Hu invariant moment feature and Hamming rule are combined to improve immune neural network for accurate sign language image matching. The experimental results show that the sign language key motion capture and recognition accuracy based on extension immune neural network has greater advantages than the classical immune algorithm.