Character Recognition Approach Involving Histogram Classification
Original Publication Date: 1979-Feb-01
Included in the Prior Art Database: 2005-Feb-19
A new approach to printed character recognition is provided which involves the pattern classification of the histogram produced by scanning the character rather than the pattern classification of the scan character itself. Pattern classification and recognition covers an extremely broad spectrum of applications and problems. Especially in the word processing business, printed character recognition (mostly typewritten characters with fixed pitch or proportional spacing) has become an important requirement, as evidenced by the new products being introduced to the market. There are several different hardware approaches (video CRT scan, magnetic scanner and digital optical array scanner), and different algorithms are being used or developed.