基于BP神经网络的人脸检测AbaBoost算法

人脸检测在日常生产和应用非常重要。本文提出了一种基于BP神经网络的AdaBoost人脸检测算法。首先，使用BP神经网络代替YCbCr高斯模型建立肤色模型。同时，针对AdaBoost算法提出了一种新的权值更新方法。在权值更新中引入阈值与样本之间的距离。另外权重有一个边界值。最后，利用BP神经网络提取图像中的肤色候选区域，并采用改进的AdaBoost算法对图像中的人脸进行精确检测。实验结果表明，利用BP神经网络和改进的AdaBoost算法的新的解决方案比现有的方法具有更高的精度。

Ababoost algorithm for face detection based on BP neural network

Face detection is very important in daily production and application. This paper proposes an AdaBoost face detection algorithm based on BP neural network. First, a BP neural network was used instead of the YCbCr-Gaussian model to build a modelofskin-color. At the same time, a new weight update method is proposed for AdaBoost algorithm. The distance between the threshold and the sample is introduced in the weight update. In addition, the weight has a threshold. Finally, BP neural network is used to extract the skin color candidate region in the image, and the improved AdaBoost algorithm is used toget the face in the image. Experimental results show that the new methodhas higher precision than existing methods.