All Things Handwritten - OCR and Beyond

Venu Govindaraju, University at Buffalo



Text is ubiquitous today on billboards, road signs, and shop banners in the outdoor scenes as well as on logos, brands, TV captions, and other written matter on everyday objects. While the field of Optical Character Recognition (OCR) has matured when dealing with conversion of scanned printed document page images to editable text, it is faced with new challenges when dealing with multilingual text images captured in the real-world by mobile cameras under uncontrolled factors of illumination, pose, and scale.

Our foray into OCR technology begins with our success in reading postal addresses and has now extended to medical forms, prescriptions, scratch notes, and annotations. We have developed successful techniques for processing handwritten and machine printed text in Roman, Indic, and Arabic scripts. In this talk we will present the state-of-the-art in the field, some of the innovative techniques developed by our group, and describe several applications in the mobile commerce space where further OCR advances can make a significant difference.

Venu Govindaraju (Bio)

Dr. Venu Govindaraju is a Distinguished Professor of Computer Science and Engineering at the State University of New York. He has authored over 370 scientific papers and supervised the doctoral dissertation of 28 students. His seminal work in handwriting recognition was at the core of the first handwritten address interpretation system used by the US Postal Service. He is a recipient of the IEEE Technical Achievement Award and a Fellow of the AAAS, ACM, IAPR, IEEE, and the SPIE.