In a previous experiment, we have studied the preference judgement of pictorial images with image experts and naive observers. We asked image experts to improve pictorial images the way they preferred. Then we showed those images to naive observers and asked them for their preferred image. We learned that, to enhance an image, an expert divides it into large zones of interest, which mainly correspond to natural colors, and that, to assess their preference judgments, naive observers principally focus on natural colors like sky, skin or grass when present.In the present work, we want to extract information about expert's behavior in enhancement and naive observers preference, directly from the digital files produced by one of the experts. We show that the segmentation process used by the expert permits to apply different correction on different objects. We use the previous work on memory colors made by Yendrikhovskij to correlate our images data with the theory of memory colors. We show that, to enhance an image, an expert moves the pixel colors toward areas of memory colors corresponding to the modified object. The expert also follows some rules while enhancing one segmented part. The corrections must be plausible inside one segment and for the whole image. The images are accepted by observers in relation with the presence of memory colors and when the treatment of the whole image seems coherent.