Mouse Manipulation Simulation based on Hand Gesture Recognition

• Main Authors: Ju-Hsuan Ou, 歐雨軒
• Other Authors: 李建誠
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/50626150787111151949

碩士 === 元智大學 === 通訊工程學系 === 98 === This thesis presents a novel vision-based human-computer interaction system. We use webcam to capture hand gesture pictures and simulate mouse manipulation. The system employs moving object detection, shadow detection, and skin color detection to obtain the hand region from captured images. Kernel density estimation (KDE) and many features on hand contour are used to recognize the mouse manipulation based on hand gestures. Additionally, the system also includes the mouse gesture recognition. It is able to recognize four kinds of mouse gestures by a mouse gesture recognition system based on hidden Markov model (HMM.)