基于肤色模型和SURF算法的人脸识别研究

人脸识别技术是机器人视觉领域研究课题之一,在军事及民用领域应用广泛。目标检测算法中,SURF算法具有精度高、鲁棒性好等优点,然而因其计算的复杂性,实时性并不高。利用人脸肤色特征,提出了一种在肤色模型的基础上结合SURF算法进行人脸识别的方案。阐述了YUV色彩空间模型以及SURF算法对特征点的检测、描述与生成的原理。并论证了YUV色彩空间与SURF算法相结合的理论原理。

Study of Face Recognition Algorithm Combined Skin Color and SURF

The technology of face recognition is a research subject in the filed of robot visual,which has been widely used in mil- itary and civilian.Belong to the target detection algorithms,despite the advantage of its high accuracy and robustness,SURF algorithm does not perform well in real-time because of its computational complexity.This paper proposes the face recogni- tion program,which is based on the skin color model and takes advantage of SURF algorithm.YUV color space mode is de- scribed in the dissertation as well as the principle of how SURF detect,describe and generate the characteristic points.Be- sides,the theory of how YUV color space combines with SURF algorithm are demonstrated in this dissertation.