Character recognition using Fourier coefficients
Amit Sharan · ThinkTech (Texas Tech University) · 1993
Character recognition has been a classical problem in the area of pattern recognition.Unconstrained handwritten characters pose a serious challenge to the researcher, and to date no algorithm has had a great deal of success in this area.The character recognition algorithms have to deal with a lot of variations in the characters, which requires great adaptability in the recognition procedure.Many approaches have been studied over the years, and several neural network implementations have been developed to classify the data.In this work, low-frequency Fourier transforms of characters are used as the features for classification.An adaptive neuro-fuzzy clustering algorithm is used to classify the features.This work discusses the basic problems faced by handwritten character recognition and the approach taken here to deal with them.