摘要
提出一种结合人脸中眼睛的灰度及形状信息对人脸检测中的眼睛进行定位的算法.先把图像分割成小图像块,根据图像块的复杂度对两眼睛的大概位置进行定位,再根据图像块的结构居中度对眼睛进行准确定位.该算法容易实现,运算量少,且与图像的平移、旋转、尺度无关.实验表明,该算法在人脸检测的眼睛定位上有较高的准确度.
The paper proposes an algorithm of eye location in face detection, combined withgrayness and shapes of eyes. Facial images are first divided into small image blocks and rough positions of eyes are located by complexity of image blocks. Finnally, the structure central measurement of image blocks is applied to the exactlocation of eyes. The algorithm is easy to implement with less amount of computation and independent of transmission, rotation and image size. Experimental results indicate that the proposed algorithm is fairly precise in eye location in face detection.
出处
《淮阴师范学院学报(自然科学版)》
CAS
2003年第3期237-240,共4页
Journal of Huaiyin Teachers College;Natural Science Edition
关键词
人脸检测
眼睛定位
复杂度计算
居中度矫正
face detection
eye location
complexity computation
correction of central measurement