摘要
许多中国画作品的图像被数字化后在因特网上展出,如何有效的识别它们并进行分类是一个值得研究的问题。提出一种基于内容的中国画识别与分类方法,通过颜色和纹理特征来表现中国画图像的可视内容。结果显示该方法能够有效的识别和分类中国画图像,其中支持向量机的分类算法可以取得最佳的分类效果。
A lot of images of traditional Chinese painting （TCP） are digitalized and exhibited on the lntemet. However, how to identify effectively and classify them are imperative problems need to be addressed. The paper proposes a content-based identification and classification system that represents the visual content of TCP images by chromatic and textural feature set. The results show that the system is capable of identifying the TCP image and classifying them based on art movements. Especially, the Support Vector Machine algorithm can lead to the most optimal classification result.
出处
《重庆科技学院学报（自然科学版）》
CAS
2009年第1期116-118,共3页
Journal of Chongqing University of Science and Technology：Natural Sciences Edition
基金
Asian-Swedish Research Links Program(Grant No.348-2005-6434)
南京市留学回国人员科技活动择优资助项目(TJ206015)
