Feature extraction and classification using a hierarchical neural network topology

Summary form only given as follows. The current trend in pattern recognition of objects of interest in an image is to use templating techniques via matched filters. The authors explore the use of a hierarchical neural network utilizing Gabor functions for feature extraction and feedforward networks for classification. The proposed combination of Gabor functions and feedforward neural networks in a hierarchical system is shown to be a viable concept for pattern recognition.<
>