摘要
在不同光照条件下同一个人脸的成像会有较大差异,这种差异的存在给人脸自动识别带来较大的限制。文中基于合理的假设,提出一种人脸图象的逐点变换模型。该模型简单易行,且能有效减少同一人脸在不同光照条件下的成像差异。在YaleB人脸图象库与AR人脸库上的对比实验表明了该模型的有效性与可行性。
Under different illumination conditions, image on an individual may vary dramatically, as a result face recognition becomes more difficult because of this factor. In this paper, a linear transformation model on face image is proposed. The model is feasible, simple and tractable in computing. Moreover, it is very effective to weaken the aftereffect of varying illumination. To illustrate the application on the model, experiments are performed on YaleB and AR face image databases.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2005年第4期846-847,892,共3页
Journal of System Simulation
基金
国家自然科学基金(60472060
60473039)
关键词
光照变化
图象正则化
图象处理
人脸识别
illumination variation
image regulation
image processing
face recognition
