A Biometrics Person Authentication Algorithm by Dynamic Feature of Facial Expression

本研究的目标是基于人脸图像，考虑人脸表情的动态特征或三维人脸几何信息，构建高质量的人脸生物特征身份认证系统。在此基础上，研究了老化人脸建模与合成、自然微笑特征提取与合成以及仅基于正面或侧面人脸快照的高性能三维人脸估计技术，为构建人脸认证系统和刑侦辅助系统做出了贡献。因此，实现了高性能，适合实际使用。

The target of this research is a construction of high quality biometrics person authentication system by face image considering a dynamic feature of facial expression or 3-d face geometry information. Moreover, a modeling and synthesis of aging face, a feature extraction and synthesis of natural smile, and high performance 3D face estimation only from frontal or profile face snapshot are studied as an element technology to contribute to construct not only a face authentication system but also a criminal investigation assistance system. As a result, high performance is achieved to be suitable for practical use.