Real-Time Chinese Sign Language Gesture Prediction Based on Surface Emg Sensors and Artificial Neural Network

The sign language recognition system is a process of acquiring hand and arm motion information through sensors and classifying the sign language. Through the sign language recognition system, deaf and mute individuals can communicate with people who can hear and speak normally by using their body language. In this paper, we propose a real-time Chinese sign language (CSL) recognition system that uses surface electromyography (sEMG) and an improved artificial neural network (ANN) classifier to recognize and predict 20 commonly used words in real time. The experimental results show that after proper preprocessing, data segmentation, feature extraction, and prediction classification, our system achieves a recognition accuracy of 91.5%. By segmenting the training set, we further significantly reduce training time without affecting the results. The results also show that about 50% of the training set is trained, and our system can achieve the desired effect.