Menu:


Outline of the talk of Prof. B. B. Chaudhuri

Title:OCR: A Pattern Recognition Success Story

Abstract and Outline:
Recognition is a high level intelligent process requiring memory with capability of matching and decision maiking. Pattern recognition is a discipline to automate the process by machine, mainly by computer. Maximum amount of patterns encountered and generated by human being are pictorial in nature. This is followed by sound patterns, such as speech and music. This talk concentrates on Pictorial pattern recognition only.

Pictorial patterns can be natural or man-made. They include images of scenes, Biomedical picture, Fingerprint pattern, ECG and other graphical pattern, handwritten and machine printed text etc. Automatic recognition of such patterns is very useful in various practical applications. Among the applications, one of the success stories is on automatic document analysis that includes Optical Character Recognition (OCR) of printed and handwritten text, Bank check reading, Signature identification and verification, Engineering drawings and diagram recognition, writer identification etc. In the current lecture this topic will be discussed, where the OCR part will be elaborated in details. Special emphasis will be put on recognition of Indian Script, namely Bangla/Assamese script.