基于特征融合与支持向量机的人脸活体检测技术

Living Face Detection Technology Based on Feature Fusion and SVM

人脸识别已广泛应用于各种场景,随之而来的人脸伪造攻击对人脸识别应用的安全性带来了极大的挑战。人脸活体检测是应对此类攻击的最佳手段。本文在分析现有人脸活体检测方式的基础上,提出了一种基于LBP与色彩饱和度的融合特征进行人脸活体检测。经实验验证,该方法准确性有了较大提升。

Face recognition has been widely used in various scenes,and the following forgery attacks have brought great challenges to the security of face recognition applications.Living face detection is the best way to deal with such attacks.Based on the analysis of the existing methods of human face detection,this paper proposes a fusion feature of LBP and color saturation for human face detection.The experimental results show that the accuracy of this method has been greatly improved.