A post processing system for global correction of Ocr generated errors

This thesis discusses the design and implementation of an OCR post processing system. The system is used to perform automatic spelling detection and correction on noisy, OCR generated text. Unlike previous post processing systems, this system works in conjunction with an inverted file database system. The initial results obtained from post processing 10,000 pages of OCR\u27ed text are encouraging. These results indicate that the use of global and local document information extracted from the inverted file system can be effectively used to correct OCR generated spelling errors