摘要： 根据视频监控应用的特点,结合视频图像的时域连续特性和人脸肤色特征,提出了一种应用于视频监控的快速人脸定位方法。该方法首先通过对称差分算法,提取运动区域;然后基于BP(back-error-propagation,误差反传)神经网络的肤色分割算法,对运动区域进行肤色检测,最后,经过进一步的候选人脸区域验证,定位图像中的人脸。实验结果表明,提出的方法实现简单、检测速度快、误检率低,适合实时视频监控系统应用。

关键词: 视频监控, 人脸定位, 对称差分, BP神经网络, 肤色分割

Abstract: In this paper, a fast face location method for video surveillance is presented, which combines the motion information and skin-color feature. Firstly, symmetrical frame difference is used to acquire the motion area. Then, the skin-color segmentation algorithm based on BP (back-error-propagation) neural network is used to extract the face candidates in the motion area. Finally, the candidate face regions are examined with the knowledge of human face. The experimental results show that the proposed algorithm has high speed and low error-detection rate, so it can be used in the real-time video surveillance system.

中图分类号: 

• TP391.4