Skip to Main Content
Neural network analysis, an important branch in data mining, has been widely used in statistical analysis, pattern recognition, image processing, biological species division and customer division. Based on division method, the paper rationally selected initial class center, dynamically regulated the number of classification during image classification, and proposed an image recognition method. In the new method, multi-scale wavelet decomposition was firstly conducted for the image to be recognized. Then fisher transformation was performed on decomposition results of different scales which were defined as decomposition vectors. Finally, image recognition was realized in the fisher transformation domain according to the minimum absolute distance or comparative distance. The new method was proved to have high correct recognition rate and excellent recognition effect.