Improved shape code based word matching for multi-script documents

Abstract

In this paper, we propose a shape code based wordimage matching (word-spotting) technique for word retrieval in multilingual documents, written in Indian languages. Each query word image to be searched is represented by a sequence of shape codes that corresponds to primitives. Then an inexact string matching technique is applied for measuring the similarity between the codes generated from the query word image and each candidate word images, obtained from the document. Based on the similarity score, we retrieve the document where the query image is found. Experimental results on Bangla, Devanagari scripts document image databases confirms the feasibility and efficiency of our proposed approach.