Abstract—Text block segmentation plays an important role in image search systems. Web pages are segmented into blocks for getting text around the image, which will later be used in retrieval process based on user query keywords. This paper presents semantic based text block segmentation for image retrieval system. Text block segmentation is performed for image indexing using semantic relevance between text blocks. Before segmentation process, web page is broken down in DOM (document object model) tree structure, where tags are used as nodes in the tree. Relatedness between text blocks are computed based on semantic relevance between terms. Semantic relevance in this paper is computed based on the hierarchical structure of the words and their hierarchical information is obtained from Wordnet dictionary. Image search by semantic relevance hence improves the accuracy of the search process

Index Terms—DOM tree, semantic relevance, text block segmentation.

The authors are with University of Technology, Yatanarpon Cyber City,Pyin O Lwin, Myanmar (e-mail: thae.thae.star@gmail.com, susuela@gmail.com).

Cite:Nyein Myint Myint Aung and Su Su Maung, "Semantic Based Text Block Segmentation Using WordNet," International Journal of Computer and Communication Engineering vol. 2, no. 5, pp. 601-604, 2013.