earticle

영어

Along with the advancement of Internet and digital technologies, more and more Chinese traditional paintings are becoming available on the Internet. As a result, computerized indexing and classification of given Chinese paintings emerge to be one of the focused research areas over recent years. As traditional Chinese paintings rely on the special drawing tools to illustrate the artistic styles, it distinguishes from western paintings in terms of strokes, contours, color and textures etc. Additionally, drawing lines play important roles in most traditional Chinese paintings. Consequently, the existing research on Western paintings is normally not applicable to traditional Chinese paintings. In addition, color-based approaches are also not applicable as traditional Chinese paintings mostly rely on gray scale texture to express their art styles and content. This thesis reports my intensive research program on computerized classification and automated learning and analyzing techniques for traditional Chinese paintings, in which a number of novel research and ideas are developed and put forward for style-based classification as well as its related theories and new concept introductions. My own novel contribution can be highlighted in machine learning, especially one-class SVM (OCSVM) based classification of traditional Chinese paintings. In this paper, the one-class SVM technology is revised to introduce a supervised learning element and arranged into a parallel OCSVM classifier. Based on the statistics features, a new concept of enforced learning has been introduced to remove the false positives at each learning cycle together with a new scheme of adaptive upgrading of decision parameters. Extensive experimental results support that the proposed new classifier achieves significant improvements in comparison with existing representative techniques.