Use of computer learning and OCR mathematical morphology in the computer stage

• Grant number: 97/07325-8
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: January 01, 1998
• End date: September 30, 2000
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computational Mathematics
• Principal Investigator: Felício Hissaaki Sakamoto
• Grantee: Felício Hissaaki Sakamoto
• Company: Carta Consultoria Ltda
• City: São Paulo

Abstract

The object of this project is to develop a study to enable the development of the use of automatic optical character recognition (OCR) in faxes sent by computer. Based on this recognition, the application must enable structured storage of information received and the carrying out of swift searches and recoveries of this information. The theory used will be based on Mathematical Morphology. This theory is useful for modeling changes between images and, for this reason, it has been used quite successfully to solve various image processing problems. By automating the morphological operators project (mapping between images), based on computer learning, it is intended to show the feasibility of developing a program able to be used automatically in offices without users having to have specific knowledge of the technique used. The expected benefits are quite broad, particularly in space saving in storing information, in enabling automatic classification of, and access to, fax contents, and in searching for subjects and specific words, something which was previously impossible with matrix (bitmap) images. (AU)