Understanding Of Visual Art Research

In recent years, studies on Art History, Computer Graphics and Digital Image Processing have been brought to the notice of researchers. The aesthetic research based on computer is named by Computational Aesthetics. Traditional studies on visual arts are depended on artists, their knowledge about the artwork and its creator. So the judgment is subjective. By contrast, Computational Aesthetics are aimed at the scientific understanding of Visual Arts and describing artworks in a objective way. The objective evaluation of Art and artworks is the research purpose of researchers. The scientific understanding of Visual Arts can be used in researches of the Art History, the authentication and the analysis of Art styles. Several techniques of studies on he scientific understanding of Visual Arts are illustrated and the advantages and disadvantages are also stated. Considering these techniques and some studies on statistical regularities of natural images, five different qualified features of styles are proposed to analyze the relations among styles, differences and similarities of them. In addition, these features are used to transfer styles to different artworks.First, techniques of the scientific understanding of Visual Arts are summarized. Hot topics of it. the trend of research and applications based on these techniques are also stated. The advantages and disadvantages of these techniques are interpreted by comparing them.Second, five different qualified features of styles are proposed based on Curvelet Transform:the multi-scale Amplitude. the distribution of coefficients, the non-stationary, the anisotropy and the correlation of coefficients. The formula and the procedure of feature extractions are represented.Third. based on these qualified features, styles of Ding Shaoguang, Xu Beihong. Qi Baishi. Claude Monet. Vincent van Gogh and Henri Matisse are analyzed respectively. The differences and similarities of styles are qualified objectively and the result demonstrates that these qualified features reflect characteristics of styles and techniques of artists in different ways.At last. these features are transferred. and the result shows the statistical regularities are consistent with the visual effect of styles.In view of the development of scientific understanding of Visual Arts, this thesis proposed five qualified features based on the Curvelet Transform and statistics are effective supplements to the studies on this area.