论文部分内容阅读
针对非均匀光照因素干扰维吾尔族人脸识别效果,提出基于CL多小波和灰度排列对(GAP)算法.首先通过归一化系统操作消弱非均匀光照对维吾尔族人脸图像的初级影响,然后利用CL多小波分解一层操作提取维吾尔族人脸图像中的非高频信息,再采取GAP算法找到维吾尔族人脸图像中固定的二维像素点差,为每个类型的维吾尔族人脸信息图像创立其对应的背景匹配模版,最后分类识别经过估算的测试样本图像和每个类型模版的匹配程度.实验结果表明,该算法在保留维吾尔族人脸图像特征的同时,极大地提高了维吾尔族人脸的识别率和运算速度,使维吾尔族人脸图像拥有了非均匀光照下良好的鲁棒性和实时性.“,”Aiming at the Uygur face recognition effect interfered by non-uniform illumination factor,an algorithm based on CL multi-wavelet and gray arrangement pair (GAP) algorithm is proposed. The normalized system operation is used to weaken the primary influence of non-uniform illumination on Uygur face image. The CL multi-wavelet decomposition one-layer operation is used to extract the non high-frequency information in Uygur face image. After that,the GAP algorithm is adopted to find the fixed two-dimensional pixel points difference in Uygur face image,and create the background matching template corresponding to the Uygur face information image of each type. The estimated test sample and matching degree of each type template are clas-sified and recognized. The experimental results show that the proposed algorithm can greatly improve the recognition rate of Uygur face image and calculation speed while reserving the image characteristics of the Uygur face,which makes the Uygur face image possess strong robustness and high real-time performance under non-uniform illumination.