An automatic yearbook style photo generation method using color grading and guide image filtering based facial skin color correction

In this paper, we propose an automatic photograph generating method for albums of large numbers of people, which gives consistency to facial skin color and background colors. In our framework, we assume photography in a simple photographic environment where some amateur cameramen take photographs indoors. In such a case, facial skin color is often distorted due to the lighting environment (e.g., light reflected from a colored rear wall), and if the photo is artificially combined with another background, the difference in color is emphasized, resulting in an unnatural synthesized result. In our technique, after roughly extracting the facial region and rectifying the distribution of the skin color in a color space, we perform correction of the color and brightness near the face of the original image so that it matches the color balance of the facial image after the color correction. Unlike conventional algorithms for color correction, the final result is attained by a correction process with a guide image. We show through experimental results that more natural results are obtained than conventional methods.