LGL-DIR: Layout graph for layout based document image retrieval

In this paper, a system for retrieving document images from digital library based on visual similarity of them is proposed. First, a graph layout created based on document extracted blocks. This graph represents document layout structure. A vector created based on path patterns to represent this graph. Then, self organizing structure is used for indexing of images in the database. This method allows us to retrieve similar images without direct comparison between the query image and database files. Experimental results on a variety of document image databases confirm the feasibility, validity, and efficiency of our proposed approach in visual based document image retrieval.