Analysis of a biologically motivated neural network for character recognition

Publisher Website
Google Scholar
Abstract
A neural network architecture for size-invariant and local shape-invariant digit recognition has been developed. The network is based on known biological data on the structure of vertebrate vision but is implemented using more con- ventional numerical methods for image feature extraction and pattern classification. The input receptor field struc- ture of the network uses Gabor function feature selection. The classification section of the network uses back-propa- gation. Using these features as neurode inputs, an imple- mentation of back-propagation on a serial machine achieved 100% accuracy when trained and tested on a sin- gle font size and style while classifying at a rate of 2 ms per character. Taking the same trained network, recogni- tion greater than 99.9% accuracy was achieved when test- ed with digits of different font sizes. A network trained on multiple font styles when tested achieved greater than 99.9% accuracy and, when tested with digits of different font sizes, achieved greater than 99.8% accuracy. These networks, trained only with good quality prototypes, rec- ognized images degraded with 15% random noise with an accuracy of 89%. In addition to raw recognition results, a study was conducted where activation distributions of cor- rect responses from the network were compared against activation distributions of incorrect responses. By estab- lishing a threshold between these two distributions, a re- ject mechanism was developed to minimize substitutional errors. This allowed substitutional errors on images de- graded with 10% random noise to be reduced from 2.08% to 0.25%.

This publication has 4 references indexed in Scilit: