Automatic Red-Eye Detection and Correction using Segmentation

J.S. Kim, D.H. Lee, D.C. Park, C.G. Jo, S.C. Kim, and B. Kang (Korea)


red-eye detection and correction, HSI color space, segmentation, border tracking


In this paper, we propose an automatic red-eye detection and correction technique using segmentation algorithm. To reduce the computational complexity and the memory requirement, the original image is divided into several segments, then the red and white mask are extracted from each segment using hue-saturation-intensity (HSI) color model. Also, we apply the border tracking instead of the conventional grassfire algorithm to group the closed red pixels. The proposed method has much less computational amount and memory usage with the almost same performance as the conventional one.

Important Links:

Go Back