Efficient Feature Extraction using DWT-DCT for Robust Face Recognition under varying Illuminations
Virendra Prasad Vishwakarma, Sahil Dalal, Varsha Sisaudia · 2018
Face recognition is an important aspect of computer vision since past many decades under uncontrolled variations such as illumination, pose and expression. In this paper, an algorithm is proposed for efficient face recognition under varying illumination by extracting robust features from the illumination normalized face images. This is performed by integrating discrete wavelet transform (DWT) with discrete Cosine transform (DCT) as the feature extraction technique. Combination of DWT and DCT is exploited so that redundancy which is not extracted by DCT alone, is firstly extracted using DWT and subsequently, the local correlation is utilized by DCT. DWT also helps in extracting the global features of the face image. The algorithm is implemented and tested over Yale, Yale B and CMU PIE face databases. As it can be seen here, promising results have been achieved by proposed approach compared to the results of the existing papers.