【摘要】 为提高人眼定位的鲁棒性、准确性和检测速度,以改善算法的适用性,提出了一种基于YCbCr彩色空间的肤色信息和基于水平灰度投影与HIS彩色空间的人眼虹膜饱和度信息相结合的算法。算法先对图像的Cb和Cr分量用改进的高斯分布模型算法进行肤色提取,通过肤色区域的分析,对人脸区域进行预检测;然后利用水平灰度投影方法定位眼眉区域,再利用人眼虹膜的饱和度信息进行精确定位。实验结果表明,该方法对人眼定位的正确率达到85%,定位时间为0.072 6 s。更多还原

【Abstract】 In order to improve the robustness of human eye positioning,accuracy and detection speed,and improve the applicability of the algorithm,a methord was proposed that combining color information based on YCbCr color space with gray level projection and human iris saturation information based on HIS color space.Color was extracted with Gaussian distribution model of improved algorithm base on Cb,Cr Component of Image,and detected face region through analysis of color regions.The gray-level projection methods to locate eyebrow regions was adopted,the human iris saturation information was used to locate eyes accurately.Experimental results show that the right accuracy of eye positioning is 85%,positioning time is 0.072 6 s.更多还原