An efficient hybrid PC-SIFT-based feature extraction technique for face recognition

In this investigation, an efficient hybrid approach involving phase congruency (PC) and shift invariant feature transform (SIFT) for face recognition is presented. The present study exploits the advantages of PC and SIFT together for the purpose of efficient feature extraction for the facial images. The effectiveness of the present work is analysed and compared using other classifiers, i.e. K-means and self-organizing map. The results of this study demonstrate that phase congruency - shift invariant feature transform is robust to expression variations and shows better performance than other comparative methods and achieves good recognition accuracy. Studies are conducted on Japanese female facial expression and Yale databases. The proposed technique has been compared with the existing techniques, and from the experiments, it is observed that the results of the proposed technique are better than the existing techniques.