A Comparative Study on Various State of the Art Face Recognition Techniques under Varying Facial Expressions

Through face we can know the emotions and feelings of a person. It can also be used to judge a person's mental aspect and psychomatic aspects. There are 5 state of the art approaches for recognizing faces under varying facial expressions. These 5 approaches are overlapping Discrete Cosine Transform (DCT), Hierarchical Dimensionality Reduction (HDR), Local and Global combined Computational Features (LGCF), Combined Statistical Moments (CSM), and Score Level Fusion Techniques (SLFT). Matlab code has been developed for all the 5 systems and tested using common set of train and test images. The train and test images are considered from standard public face databases ATT, JAFFE, and FEI. The key contribution of this article is, we have developed and analyzed the 5 state of the art approaches for recognizing faces under varying facial expressions using a common set of train and test images. This evaluation gives us the exact face recognition rates of the 5 systems under varying facial expressions. The face recognition rate of overlap DCT on ATT database was 95% and FEI 99% which was better than HDR, LGCM, CSM and SLFT. But the face recognition rate of CSM on JAFE database, which contains major facial expression variations, was 100% which was better than overlap DCT, HDR, LGCM, and SLFT.