Directional Edge Coding For Facial Expression Recognition System
K. Venkata Subbareddy, P. Sandhya, B. Pavani · 2023
The local appearance-based texture descriptors used in facial expression recognition are only partially accurate. The main reason is the existence of noise-induced distortions and weak edges. Hence, this paper proposes a new local texture descriptor called as Directional Edge Coding based Facial Expression Recognition System (DEC-FERS). DEC-FERS considers the neighbouring pixels support for determining facial expression semantics including edges, corners, lines, and curves. DEC-FERS extracts weaker edge responses through edge detection masks and discards them after encoding only more robust edge responses. Robinson Compass Mask and Kirsch Compass Mask are used to extract the edge responses from facial images. JAFFE and OU-FER databases are used for the experimental validation and the performance is assessed through recognition accuracy.