A study on improvements for face recognition

• Main Authors: Yu-Hao Chang, 張宇豪
• Other Authors: 林聰武
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/s9k6gt

碩士 === 東吳大學 === 資訊科學系 === 96 === In this paper, we present a scheme of face recognition, which only use the face features like eyes, eyebrows, nose and mouth to find the angle information. And through the Hausdorff distance with mouth feature to compensate the various expressions. Experiments on the YALE face database, we gather the feature place to reduce the dimension and through the illumination and face expressions compensation method to achieve higher recognition rate (95.3%) than some methods. Besides enhancing the recognition, our proposed method reduces the computation time is owing to the simpler algorithm.