ALBP识别方法及其在聋人视觉辅助系统中的应用  

Identification Method of ALBP and its Application in Aid System for Deaf Vision

作　　者：杨小义[1] 邓新梅[2] 

机构地区：[1]重庆师范大学教育科学学院,重庆401331 [2]重庆科技学院图书馆,重庆401331

出　　处：《重庆理工大学学报（自然科学）》2015年第8期94-98,共5页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金青年基金资助项目"基于度量学习的特征提取方法研究"(1010200720100098)

摘　　要：针对LBP视觉识别算法难以消除光照变化强烈导致的图像非线性扭曲的问题,提出了一种改进的ALBP识别方法。该方法首先对不同光照下人脸的图像进行预处理,然后将像素间的对比度值映射到某个指定区域以抽取出人脸图像的LBP特征值,最后再将该特征值进行融合计算,即可得到在复杂光照条件下比较真实的人脸图像。实验仿真对比研究结果验证了ALBP方法对人脸图像识别的合理性与可行性。研究结果表明,ALBP识别方法在复杂光照下可比较好地消除图像的非线性扭曲。Aimed at the puzzle of being difficult to eliminate the image non-linearity angulation resulted in stong illumination variations by using vision identification algorithm of LBP, the paper presented a sort of improved identification method of ALBP. It firstly made the pretreatment for human face image under different illumination and made the contrast ratio value among pixels to be mapped into a certain specified region so as to extract the feature value of LBP in human face image, and then carried out the fusion computing for the feature value, and finally it could obtain the more true human face image under the condition of complex illumination variations. The comparative study of experiment simulation demonstrates the rationality and feasibility for identification of human face image by the method of ALBP. The study result shows that it can better eliminate the non-linearity angulation of the image under complex illumination variations.

关 键 词：视觉识别 聋人交流 图像非线性扭曲 识别算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]