Facial Image datasets

Description:
This repository contains the datasets used in the study titled “MAS Machine: Multi-kernel Attention Spatio Machine.” The experiments were conducted using three widely recognized benchmark datasets for facial recognition under varying conditions:

• AR Face Dataset – includes variations in facial expressions, illumination, and occlusions such as sunglasses and scarves.
• Extended Yale B Dataset – contains images captured under extreme illumination conditions, making it suitable for evaluating robustness to lighting variations.
• CK+ (Cohn-Kanade) Dataset – widely used for facial expression analysis, containing diverse emotional expressions and pose variations.

These datasets were utilized to evaluate the robustness and effectiveness of the proposed multi-kernel attention-based framework under real-world conditions, including variations in lighting, expressions, and occlusions.