Twelve Directional Feature Extraction for Handwritten English Character Recognition
Sumedha B. Hallale, Geeta Salunke · 2013
Directional features have been successfully used fo r the recognition of both machine printed as well as h andwritten characters. Selection of feature extraction method is probably the single most important factor in achieving high perf ormance in pattern recognition. In this paper, twelve direction al features are used for the recognition of handwritten English alpha bets and numerals. The properties of similarity measure are analysed with directional pattern matching. Then the comparison is made between recognition rate of conventional and twelve d irectional feature extraction techniques. The experiment shows that directional feature extraction techniques are bette r than conventional one.