Intelligent OCR processing

Optical Character Recognition (OCR) has become a highly demanded information transfer technology in recent years. This demand has been driven by the increasing needs for information sharing and office automation, and by the increasing accessibility to large‐scale, fast, and powerful computer resources. A problem of current OCR technology is that texts produced by the state‐of‐the‐art OCR software contain an unacceptable frequency of errors. This prevents the OCR technology from being efficiently used for vast‐volume information transfer or daily office operation applications. To correct these errors in a conventional way requires a significant amount of costly human‐machine interaction. In this article, we identify and classify the types and distributions of optical recognition errors. We propose a novel post‐processing strategy, based on machine learning techniques, to correct errors resulted from unrecognized or misrecognized characters during the recognition process. By applying this strategy, the accuracy of recognition can be significantly improved, and the human interaction required can be dramatically reduced. Experimental results indicate that, in a typical environment, about 46% of total errors can be corrected automatically (i.e., without human interference), with an accuracy of 91%. © 1992 John Wiley & Sons, Inc.