【摘要】 近年来,人脸检测识别方法得到飞速发展,基于ASM特征的识别方法因具有定位精度高而得到广泛应用。然而,当相邻两个特征点位于不同器官时,使用该方法时特征点间会产生连带效应。本文提出基于特征绑定的ASM人脸检测算法,克服了传统ASM方法存在的连带效应。由于使用该方法不需要直接检测完整的人脸,因此其对人脸偏转、部分人脸遮挡、用户佩戴眼镜等情况有较高的准确性和鲁棒性。更多还原

【Abstract】 In recent years, face detection and recognition methods have been developed rapidly. Recognition methods based on ASM features are widely used because of their high positioning accuracy. However, when two adjacent feature points are located in different organs, there will be joint effect between feature points when using this method.This paper proposed an ASM face detection algorithm based on feature binding, which overcame the joint effect of traditional ASM. Because this method does not need to detect the whole face directly, it has high accuracy and robustness for face deflection, partial face occlusion, users wearing glasses and so on.更多还原