Skip to Main Content
With rapid advancement in technology, numerous applications are required, such as for face and gesture recognition. However, various methods previous researchers have developed and presented suffer from limitations. Therefore, this study proposes an FPGA-based gesture recognition system by rewriting the largest computational complexity of optical flow to perform parallel processing architecture, and presents algorithm designs for the median filter, skin color detection, optical flow for hand detection, and the Kalman filter for tracking. The advantage that separates the proposed method from previous approaches is its foundation on optical flow technology, which renders the hand shape unlimited in gesture recognition. The results indicate that the cost-effective FPGA system under a low clock rate may be used to realize the proposed system. Using the FPGA system, gesture detection and recognition can achieve 30 frames per second, and the system software can subsequently schedule all tasks during processing. This study provides a simple background in an applied environment, which can be used for consumer applications, such as entertainment or medical applications, with non-touch control equipment.