基于肤色检测和Hough变换的彩色图像人眼定位  

EYE Positioning in Color Image Based on Complexion Measuring and Hough Transformation

作　　者：陈佳佳[1] 黄穗[1] 

机构地区：[1]暨南大学信息与技术学院,广州510632

出　　处：《微计算机信息》2010年第5期208-209,168,共3页Control & Automation

摘　　要：提出一种基于肤色检测和Hough变换两步快速定位人眼的方法。通过YCbCr肤色模型对图像肤色区域进行检测,同时进行Sobel边缘提取,然后用Hough变换的圆检测方法对二值化图像进行圆检测,从而确定双眼的精确位置。实验结果表明该算法受噪声和边缘间断影响小,具有较高的定位准确率。同时具有很好的容错性和健壮性。Based on complexion measuring and Hough transformation, a two-step method which is able to rapidly position human eyes has been advanced. Having detected image complexion area through YCbCr complexion model and drawed edge with Sobel to the image, round detection method of Hough transformation is applied to measure the round to two value image for positioning eyes accurately. It can be demonstrated in the experiment results that the proposed algorithm, which has low impact by noise and the edge of intermittent, would provide higher nositioning accuracv with good tolerance and robustness.

关 键 词：人脸检测 肤色检测 眼睛定位 HOUGH变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]