In this paper, we propose a novel framework for the web-based retrieval of Chinese calligraphic manuscript images which includes two main components: 1). A Shape-Similarity(SS)-based method which is to effectively support a retrieval over large Chinese calligraphic manuscript databases [19]. In this retrieval method, shapes of calligraphic characters are represented by their approximate contour points extracted from the character images; 2). To speed up the retrieval efficiency, we then propose a Composite-Distance-Tree(CD-Tree)-based high-dimensional indexing scheme for it. Comprehensive experiments are conducted to testify the effectiveness and efficiency of our proposed retrieval and indexing methods respectively.
展开▼