Face Recognition Using Active Shape Model

• Main Authors: Chun-Ching Wang, 王俊欽
• Other Authors: Chaur-Heh Hsieh
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/96690707349396253464

碩士 === 義守大學 === 資訊工程學系 === 92 === Active Shape Model（ASM） is a flexible and deformable model that can be used to represent complex object. It can be easily applied to medical science, face and industrial recognition. In traditional ASM, the labeling of landmarks is manual which is very complicated process. Generally, the ASM need a large number of landmark to represent the detail feature, sufficiently. Therefore, the computation complexty is very high and difficult to apply the practical applications. We compare the facial shape with Hausdorff Distance. The advantage of Hausdorff Distance is that the operation is simple, can reduce the computation complexity significatly, the other advantage is not limited in point-to-point comparison. Experiment results show that the new method achieves very good performance than that of traditional ASM. In this thesis, we developed a new ASM which can be generated automatically. We focus on most significant feature in face, so fifteen landmark is sufficient to represent the feature of face.