Real-time hand gesture recognition system based on Q6455 DSP board

This paper presents detailed description of a real-time hand gesture recognition system using embedded DSP board and image processing approaches. Such a system which can identify hand postures and dynamic gestures has manifold potential applications range from sign languages to human computer interaction. We use Q6455 DSP board based on 4 TI-TMS320C6455 DSPs as the computational unit. This versatile signal processing module has powerful computing and interconnection capability. To recognize hand postures, algorithm based on skin color segmentation and geometric invariant feature has been used. During identify dynamic gestures, optical flow tracking approach and direction encoding are adopted. The implementation of high reliable algorithm on DSP board keeps the system robust and efficient. Experimental results show that the proposed system performs well in recognizing hand postures and dynamic gestures real timely. The accuracy and scalability of this system are also soundly proved.