MMX-Accelerated Real-Time Hand Tracking System

We describe a system for tracking real-time hand gestures captured by a cheap web camera and a standard Intel Pentium based personal computer with no specialized image processing hardware. To attain the necessary processing speed, the system exploits the Multi-Media Instruction set(MMX) extensions of the Intel Pentium chip family through software including. the Microsoft DirectX SDK and the Intel Image Processing and Open Source Computer Vision (OpenCV) libraries. The system is based on the Camshift algorithm (from OpenCV) and the compound constant acceleration Kalman filter algorithms. Tracking is robust and efficient and can track hand motion at 30 fps.