中文摘要：

近年来，人脸检测作为一个热门的研究领域，诸多有效的检测方法层出不穷。在检测结果和检测速度方面较原来的检测方法都有很大的提升。在众多的检测算法中，FBLBP（Hexible Block—basedLocal Binary Paterns）算法在检测速度和准确度方面均有优异的表现。但在多人脸环境，单一的人脸检测结果还不是很理想。在FBLBP算法的基础上融合肤色检测，即对图像进行肤色预处理，将图像中的肤色区域优先提取出来，然后使用改进后的FBLBP算法在肤色区域优先进行人脸检测，从而使人脸检测的速度和准确度有更大的提高。采用Madab进行编程实验，使用收集的图片对FBLBP算法、融合肤色和改进后的FBLBP算法进行对比。实验结果表明，融合肤色和改进的FBLBP算法的速度和准确性均有所提高。

英文摘要：

As a hot area of research, a number of effective face detection methods have emerged in recent years. Compared with the state of art methods, the new face detection methods perform better in the speed and accuracy. Among these methods, FBLBP （Flexible Block -based Local Binary Patterns） algorithm has excellent performance. Only to use single method,in the multi-face environment, cannot reach the desired effect. The skin color detection and FBLBP algorithm is combined. Skin color preprocess is carried out to extract the skin color area and then the improved FBLBP algorithm is used to detect the face in the area extracted, so that the speed and accuracy of face detection has greater improvement. For experimental verification, the Matlab programming is adopted to compare the FBLBP and improved FBLBP. As a result,the improved one has better performance in speed and accuracy.