A new robust face detection in color images

In this paper, we propose a novel approach for robust skin segmentation and face similarity measurement. The proposed skin segmentation is a method for integrating the chrominance components of any color model. The goal of the skin detection method is to select the appropriate color model for verifying the skin pixels under different lighting conditions and various types of skin color. An enhanced Hausdorff distance, called the robust automatic minimum Hausdorff distance (RAMHD), is used to measure the similarity between the face edge and an elliptical model in the skin area. This method is very robust to occlusions of the face edge. Finally, the results of face similarity measurements are improved by updating the elliptical model. We show the performance of skin segmentation and face similarity measurements with real images.