基于肤色模型和径向基函数网络的脸部检测

驾驶员疲劳状态检测一般采用对人眼的闭合度进行计算,若实现对人眼的闭合度计算首先是对人脸的正确快速检测,针对驾驶室的特定环境,本文研究一种基于肤色模型和径向基函数网络为基础的快速人脸检测算法,该算法首先对输入图像进行RGB和YCbCr颜色空间的转换,其次建立相关的肤色模型,实现人脸区域的粗定位,然后结合径向基函数网络对输入的图像进行训练,这样就可以根据训练的结果判断是否是肤色,从而实现人脸检测。仿真结果表明,所研究的算法较好的提高了强光下人脸的正确检测,为驾驶员疲劳驾驶的研究奠定前期基础。

Face Detection Based on Skin Color Model and Radial Basis Function Network

The calculation of the closure of human eyes is commonly adopted to detect driver fatigue. In order to realize human eyes closure calculation, correct and rapid detection of human face is accomplished firstly, for the specific environment of cabs, this paper proposes a fast face detection algorithm based on skin color model and radial basis function network, which makes input image carry out RGB and YCbCr color space conversion, then establishes relevant skin model to achieve the coarse positioning of face region, finally, combines radial basis function network to train input image, so that whether it is the skin color is determined according to the training results, and the detection on face is finished. Simulation results show that the algorithm improves the human face correct detection un-der strong light, laying a foundation for drivers’ fatigue driving research.