AI-Powered Classification of Artistic Attributes

In view of the current situation in image classification research where algorithms for classification based on the art-style attributes of images are still relatively rare, an automatic image classification system based on art attributes has been implemented. This system mainly involves several typical art-style images such as photographic works, traditional Chinese paintings, watercolor paintings, sketches, and oil paintings. The system uses the Support Vector Machine (SVM) as a classifier and adopts a hierarchical classification method. A specific SVM binary-tree multi-class classification algorithm for the classification of art-attribute images is proposed. Then, through the analysis of the art-style characteristics of various types of images, representative, well-discriminative, and easy-to-calculate features are respectively extracted. Finally, experimental analyses are carried out for the characteristics of each level of classification and the overall characteristics of the classifier. The experimental results show that the system has good classification performance.