A novel character-recognition method based on Gabor transform

In this paper, we give a novel character recognition method. This method includes three steps: preprocessing, feature extraction and recognition. In preprocessing, we resolve slant and distortion of character images by a minimal moment of inertia and rotation algorithm. And we effectively detect a character´s edge using Canny arithmetic operators. Then we present a novel and effective feature extraction method based on the Gabor transform. Different from other existing means, this method computes ratios of maximum from the Gabor transform outputs of character´s edge at rows and columns respectively. The feature vector constructed by maximum ratios can exhibit desirable characteristics of local statistic and orientation selectivity. We test this method on 785 character images which are from USPS and carry out the recognition work by a 3-layer BP neural network. Experiments indicate that this recognition method can achieve a recognition accuracy as high as 96.5% for these characters.