Motion trajectory based human face and hands tracking for sign language recognition

Abstract

The real life communication is not possible without interaction which is consist of text, voice or visual expressions. The communication among the deaf and dumb people is carried by text and visual expressions. Lacking of proper copra and its feature representation makes the sign communication a hot issues in machine learning research. In this work, it has been proposed a sign language recognition scheme for hearing impaired people. It has been computed trajectory by Cam Shift algorithm from the face and hands motion. Hidden Markov Model is used to recognize the signs. The problem of automated sign language recognition in video sequences can be divided into many inter-dependent modules. These include hand and face detection, hand tracking, finger tracking, feature extraction and gesture recognition. This research achieved 97% accuracy for single and double hand gesture and 70.74% for overall signs recognition in which 78.12% for double hand and 67.74% is for single hand sign recognition.