| Summary: | 碩士 === 國立臺北科技大學 === 電腦通訊與控制研究所 === 90 === To increase convenience of “access control”, the distinctness of human face feature is useful. The difficulty to recognize the human face is how to find the face location of the people in the complex environment. In this study, we have proposed an approach that can find the invariant face features from the different poses. We utilize color and smooth properties of the skin to segment human face from complex environment immediately. Using the approach, we have developed a face recognition system. The system consists of two stages, one is the learning stage and the other is the recognition stage. In learning stage, we store the detected invariant features of the people to one database. Finally we use these features to perform the face recognition through gray matching. We can detect the most similar person from the database as candidate. In the thesis, the several experimental results are also demonstrated to verify the proposed approach is feasible for face recognition.
|
|---|
