A Word Image Coding Technique and its Applications in Information Retrieval from Imaged Documents

With the need of current fast evolving digital libraries, an increasing amount of documents are being digitized into an electronic format for easy archival and dissemination purposes. Thus Document Image Retrieval (DIR), as part of information retrieval (IR) paradigm, is receiving attentions among the IR communities in recent years. This paper presents two DIR applications based on a word image coding technique to extract features from each word image object and represent them using feature code strings for comparison. The first application is a web-based retrieval system that retrieves document images online from digital libraries based on a set of input query words. The second one is a plug-in search tool embedded in Acrobat Reader that performs word spotting within the opened document images and marks the matching words explicitly in the document. Both applications achieve good precision and recall according to our experiments on document images such as students’ theses provided by our university digital library.