Skin Color Based Lightness Correction Method for Digital Color Images

摘要

We developed a new method for correcting lightness in images in response to the fact that people are most critical about the color of human skin in photographs, and other forms of image reproduction. Conventional methods that are based on color distribution of an image can finely correct the lightness of most of the image, but cannot always finely correct the lightness of the skin. An image with an inappropriate skin lightness is not always evaluated highly, even if the lightness of the entire image is acceptable, because we pay special attention to skin color. We focused on the fact that the accuracy required to reproduce the lightness of the skin and an entire image is different. We think that deviation from the optimal lightness of skin color seriously affects perception, while deviation of an entire image has much less effect. Therefore, images with people can be further enhanced by adjusting lightness according to the difference between the actual lightness and the optimal lightness of skin color after the lightness has been corrected on the basis of color distribution. We developed a lightness correction method based on this phenomenon. We evaluated our method by visual assessment. The results reveal that the performance of our method is superior to conventional methods.