Feature extraction for character recognition using Gabor-type filters implemented by cellular neural networks
V. Tavşanoglu, Ertuğrul Saatçı · 2002
This paper proposes an approach for feature extraction using a CNN Gabor filter and an orientation map. We use a set of handwritten characters for testing the complete system. The frequency response of the CNN Gabor-type filter and the filter output are studied for different values of the filter parameters.