A neural network based edge detector

An approach to the edge detection problem based on the nonlinear mapping and generalization capabilities of multilayer feed forward neural networks is proposed. The task of edge detection is broken into two parts, i.e., mapping typical gray levels in primitive small image blocks (e.g., 3×3 windows) to their corresponding most likely edge patterns using a simple neural network, and combining this locally derived information (including presence, orientation and strength of edge) in a consistent way. Some edge detection experiments based on this scheme are provided. The suggested scheme, because of its parallel structure, is fast and can be easily implemented using analog VLSI hardware