Edge Detection Based on Image Features and LVQl Neural Network

摘要

The paper for the first time brings forward a new method of edge detection based on LVQl (Learning Vector Quantization) neural network and the matchable feature vector with better anti-noise. Traditional edge detection arithmetic operators produce a lot of false edge while detecting the image stained by noise, so we bring forward a feature vector with better anti-noise that is the import of the LVQl neural network. The LVQl network is a kind of mixed neural network that adopts both the competitive learning without supervision and the learning with supervision.