ABSTRACT
The pulse-coupled neuron, which is significantly different from the conventional artificial neuron, is a result of recent research conducted on the visual cortex of cats and monkeys. Pulse-coupled neural networks (PCNNs) are modeled to capture the essence of recent understanding of image interpretation processes in biological neural systems. Our study indicates that the PCNN is capable of image smoothing, image segmentation and feature extraction. The PCNN reduces noise in digital images better than traditional smoothing techniques. As an image segmenter the PCNN performs well even when the intensity varies significantly within regions, and adjacent regions have overlapping intensity ranges. This article describes the theory, design and implementation of an image segmentation/detection system based on the PCNN.
