A Color Harmony Measure Model with Shape Information

In the physical world a visible object has color and shape features synchronously. Numerous color harmony principles and measure models were presented on the basis of psychological experiments, typical color solids, Birkhoff's aesthetic measure, or information theory. However, none of them integrally took account of shape information in measuring color harmony. Based on Lewin's psychological field and general physical force field theory, this paper introduces a new concept called shape-color field to express the interactions between elements of an image. According to finite element analysis (FEA) theory, the harmony score of an image can be calculated with the pixel color and topology information in an image matrix on the basis of harmony model of two-color combinations. The color harmony measure model proposed in this paper was applied to a set of paintings, and the results were compared with previous researches.