Extraction of Artists' Color Features of Art Paintings and its Application to Color Image Correction

摘要

This paper proposes a method to extract artists' color features of art paintings, and correct the color-mismatch images of paintings based on artists' color features. First, we describe a standard image database consisting of famous oil paintings we captured directly, and an algorithm for extracting artists' color features based on the color distribution of the standard image data for each artist. The color distribution is analyzed by PCA and described with an ellipsoid to represent the standard color features for the artist. It is suggested that the color features for the respective artists are similar to the human visual assessment of their paintings. Next, the color correction is based on the coordinate transformation of pixel values in a color-mismatch image so that the color features of the mismatched image are fitted to the artist's color features in the standard image database. We present a correction algorithm using sRGB and CIELAB color spaces. Experiments are performed using samples of color-mismatch images of the famous oil paintings.