In this paper, we propose a method for analyzing color features of famous painters representing impressionism. In the analysis, we used image data captured from real paintings at museums ourselves by using a color-calibrated camera. Therefore, we describe comprehensively a series of analysis methods from colorimetric image acquisition to color feature extraction of the art works for individual painters. First, we describe how to capture the colorimetric image data of art paintings by a digital camera in museums. We note that as the museums have different lighting environments, we standardize the captured images at different museums to create a standard image database. Next, an analysis method is presented in which the image size is made uniform, a set of works for each painter is collected, and the image data are plotted in a uniform color space.

　The overall characteristics of paints used by each painter are obtained by principal component analysis of the entire color distribution. Furthermore, three color attributes of lightness, saturation, and hue are used to analyze each painter's work. By developing a method to analyze each attribute by equally dividing the number of pixels on the attribute axis, we show that the division point and the division interval are effective for color feature extraction.