Searching handwritten documents is a relatively unexplored frontier for documents in any language. Traditional approaches use either image-based or text-based techniques. This paper describes a framework for versatile search where the query can be either text or image, and the retrieval method fuses text and image retrieval methods. A UNICODE and an image query are maintained throughout the search, with the results being combined by a neural network. Preliminary results show positive results that can be further improved by refining the component pieces of the framework (text transcription and image search). © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Srihari, S. N., Ball, G. R., & Srinivasan, H. (2008). Versatile search of scanned Arabic handwriting. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4768 LNCS, pp. 57–69). https://doi.org/10.1007/978-3-540-78199-8_4
Mendeley helps you to discover research relevant for your work.