Feature selection using Binary-ABC algorithm for DWT-based face recognition

Face recognition is non-invasive due to various challenges like illumination variation, pose variation and limitation of 2D images from most of the image capturing technologies. In this paper three novel techniques are proposed namely Binary Artificial Bee Colony (BABC), horizontal feature extraction and feature gallery expansion. BABC is a binary version of Artificial Bee Colony (ABC) which is employed as feature selection technique for efficient reduction in selected features. It optimally selects the features from the feature vector space. Horizontal feature extraction is used for extracting unique features for face images. Feature gallery expansion is employed to increase the feature galley size for better recognition. Experimental results on two standard face databases namely LFW and CAS-PEAL indicates the consistency of the proposed techniques and prominent enhancement in face recognition.