节点文献
大角度平面旋转人眼定位方法
An eye location method of big angle rotation face at normal illumination
【摘要】 提出一种解决平面大旋转角度人脸眼睛精确定位算法,首先根据圆周旋转不变性,建立基于圆周块子矩阵的标准人脸旋转模板,使对旋转角度的判定转化为对圆周块子矩阵移位距离的估计;然后用基于人脸结构关系的方法粗定位时,提出自适应的最优阈值方法去除噪声点,减少匹配误差,减少运算量,提高运算速度;最后对候选区域提取gabor特征,进行精确定位。在FRVT库上的实验结果表明,该算法适用于360°内的平面旋转的人眼定位,算法简单,运算速度快,准确率可达到87.67%。
【Abstract】 In this paper we propose a new eye location algorithm of big angle rotation face based on circle-block by which the estimation to angle rotation is transformed into computing the moving distance of circle-block matrix firstly;then a adaptive binarization threshold method is given to remove noise points in eyes coarse location by the facial structure;Finally,gabor features of candidate facial regions are extracted to locate two eyes accurately.The computational complexity of this algorithm is low.The experiment results show that the algorithm can be used to locate eyes in any rotation angle and we can achieve an average accuracy of 87.67% on FRVT database.
【Key words】 circle-block matrix; eye location; face structure; adaptive threshold; rotation face; face recognition;
- 【文献出处】 光电子.激光 ,Journal of Optoelectronics.Laser , 编辑部邮箱 ,2009年01期
- 【分类号】TP391.41
- 【被引频次】2
- 【下载频次】228