Facial expression representation and classification using LBP, 2DPCA and their combination

Facial Expression analysis is an interesting and challenging problem and has applications in many areas such as human computer interaction and robotics. Deriving an effective facial representation from original face images is an important step for successful facial expression recognition. In this paper, we are evaluating 2DPCA and LBP+2DPCA for facial representation. The three stages of facial expression recognition are pre-processing, features extraction and classification. Many researchers´ uses face detection as a pre-processing step which improves the accuracy but also increases the time complexity of the system. To reduce the computational complexity we propose to apply 2DPCA on input images directly. Our system has achieved high accuracy as well as very low time complexity. This system is suitable for real time applications. To improve the accuracy of the system we have applied 2DPCA on LBP images in place of original images.. The comparative analysis of both methods is done on the basis of their recognition accuracy and time complexity through experimental results. The proposed system has achieved the recognition rate of 95.12% for 2DPCA and 95.83 % for LBP+2DPCA. The time required to recognize an expression for 2DPCA is very less as compared to other contemporary methods.