基于反射感知模型的人脸识别算法研究

针对复杂的光照条件人脸识别问题,提出一种基于反射感知模型的光照补偿算法。通过模拟人类视觉成像系统,从成像S估计出照度L,剔除并分解出入射度R,从而消除不同强度、不同方向光源带来的光照不均的影响,达到光照补偿的目的;在此基础上,通过提取人脸二维Gabor特征,采用均匀下采样结合主成分分析(Principal Component Analysis,PCA)和Fisher线性判别分析(Fisher Linear Discriminant,FLD)对人脸Gabor特征进行两重降维处理,并结合SVM(Support Vector Machine)实现人脸的分类识别。通过实验证明所提出的算法能够有效提升复杂光照条件下的识别精度。

Research on Face Recognition Algorithm Based on Reflective Perception Model

This paper proposes a light compensation algorithm based on reflective perception model. By simulating human visual imaging system, the illumination L] is estimated from imaging S], the incident degree R] is removed and decomposed, so as to eliminate the different intensity of light from different directions to bring the impact of uneven lighting, and achieve the purpose of light compensation. By extracting the human face two-dimensional Gabor feature, using uniform down-sampling, combined with Principal Component Analysis（PCA） and Fisher Linear Discriminant（FLD） analysis, the two dimensions are reduced and combined with Support Vector Machine（SVM） to realize the classification and recognition of face. The experimental results show that the algorithm proposed in this paper can effectively improve the recognition accuracy under complex lighting conditions.