一种彩色证件照片的眼角和嘴角定位方法

A Method of Eyes and Mouth Corner Location in Identity Photographs

基于彩色标准学生证件,提出一种简单,有效的眼角和嘴角定位方法。首先,对图像进行亮度调整,颜色空间转换等预处理,再根据YCbCr空间中人脸肤色的聚类特性,分割出肤色区域,并利用数学形态学操作进行处理,最后,利用人脸的结构特征和几何分布规则,在小范围内准确定位眼角和嘴角的位置。实验结果表明,对于有噪声的,存在一定表情的图像,该方法具有较高的定位准确率,并且算法实现简单。

This paper presents a simple and effective method for locating eyes and mouth corners based on standard students＇ color documents. Firstly, executing brightness adjustment and color space conversion for images. Then, segmen- ting the image according to the skin--color＇s clustering characteristic in the YCbCr color space, and dealing with the detected region through the mathematical morphology operations. Finally, accurately positioning the location of eyes and mouth corners in a small range using the structural characteristics and geometric distribution rules of the face. Experimental results show that this method achieves better location rate for images with noise and a certain expression, and it is simple compared to traditional methods.