Title :
The use of steerable filters for hand-written character recognition
Author :
Tufan, Emir ; Tavsanoglu, Vedat
Author_Institution :
Dept. of Electron. Eng., Univ. of Istanbul Avcilar, Istanbul, Turkey
Abstract :
We develop a system for image recognition using steer-able filters and neural networks where the main role of steerable filters is in the pre-processing of the image. This process enables the characterisation of the image by the local dominant orientation rather than global image. The Kohonen´s self organising feature map is used for the recognition task.
Keywords :
handwritten character recognition; image filtering; image recognition; self-organising feature maps; Kohonen self organising feature map; hand-written character recognition; image recognition; local dominant orientation; neural networks; steerable filters; Biological neural networks; Image edge detection; Interpolation; Neurons; Transforms;
Conference_Titel :
Signal Processing Conference (EUSIPCO 1998), 9th European
Conference_Location :
Rhodes
Print_ISBN :
978-960-7620-06-4
