Study of Image-Based Expression Recognition Techniques on Three Recent Spontaneous Databases

Recent work in the recognition of naturalistic expressions, which is also known as spontaneous facial expressions recognition, has attracted researchers' attention due to its importance in different behavioural and clinical applications. The main design challenges in the area of emotion computing for automatic recognition of spontaneous facial expression are the face pose, capture distance, illumination variation, head rotation, and occlusion. Therefore, designing a robust system to mitigate these challenges is essential for real-time applications. In this paper, we present a comparison of the performance of image-based expression recognition in three types of recent spontaneous databases by using principles of sparse representation theory. The three spontaneous databases are the Video Database of Moving Faces and People (VDMFP), MMI Facial Expression Database and Belfast Induced Natural Emotion Database each having different challenges and the study aims to show which types of spontaneous conditions are more challenging in terms of system accuracy. We demonstrate through the straightforward analysis of results in terms of the error rates which aforesaid spontaneous database is more challenging. Then, we compare the use of difference images, in order for the creation of the decomposition of expressive images. The difference images emphasize the expressive areas in the face while eliminating the irrelevant parts; in this way, the identity of the facial image is removed and the identity-independent expression recognition problem is addressed.