基于级联神经网络的人脸检测方法的研究

CASCADING NEURAL NETWORKS BASED HUMAN FACE DETECTION 

提出一种基于颜色信息与多级神经网络的人脸检测方法,通过把基于不同颜色分量的多级神经网络级联起来,采用彩色图像的亮度分量和色度分量作为级联神经网络的分类特征,从而有效提高了人脸检测的正确率,同时也加快了人脸检测速度.

A new face detection method based on color information and multi stage neural networks was proposed. The system utilizes brightness and chrominance components of color images as the features of neural netwrok classifiers, and cascades multiple color components based neural networks. Experimental results show that this method can detect human faces effectively and fast.