Facial Identification using Haar Cascading with BRISK

A swift and efficient facial identification system is a well-known discipline of computer vision applications and thus forms a pivotal part of image processing. Despite the advantages of existing methods, there is still a great demand for modifications in these algorithms in order to close to lacks of proposed methods. In this proposed work, the image is preprocessed with Contrast-limited Adaptive Histogram Equalization and faces are detected with Haar-Cascading (Viola-Jones algorithm). Then face identification is done using BRISK (Binary Robust Invariant Scalable Key-points) descriptor. Experimental results demonstrate that the proposed methodology achieves better facial identification even under various challenging conditions compared with the existing BRISK.