Texture based multi-view face recognition in noisy images using BRINE feature

The human face is a standout amongst the most famous trademark which might be employed in the biometric security framework to recognize a person. Face recognition in images is a fast growing furthermore, challenging issue in computer vision with its fundamental application in law enforcement and justice solutions, airdrome security, money services, scene analysis and access management etc. However, the crucial task remains to provide reliable recognition accuracy under multi-view and noisy conditions. This paper presents a novel and effective texture descriptor BRINE (Binary Rotation Invariant and Noise Tolerant Euclidean metric feature) for multi-view face recognition. The proposed system employs PCA method to achieve dimensionality reduction on the BRINE feature descriptors and thus reduce time complexity. Performance of the proposed strategy is assessed using feature, different distance measurement, and different noise based schemes with AT& T, Yale B datasets. The experiment results on these datasets exhibit that BRINE feature vector along with PCA and SVM significantly improves the performance.