Abstract:An approach is developed to detect eyes on a complex background.First, the skin color is separated from the background using a color-grouping algorithm. With the resulting skin-color map, all potential irises are located using the Hough transform and geometric constraints for the eyes. Eye regions are then extracted based on the positions of the potential irises. Each of the eye-regions (E1) is projected onto the eigen space and then reconstructed as E2. The final eye positions are determined by measuring the correction between E1 and E2. Experiments have been performed using 980 images taken in different environments. The results show that the presented method is quite robust and efficient for various poses, ages and illumination conditions.

Detecting Facial Features Based on Color Segmentation and KL Transform