Abstract
When images have complex features, the single neural network model can't often process them with enough accuracy. This report proposes image processing methods (compression and segmentation) by using multiple multi-layered neural networks. Moreover, we show the characteristics of the proposed methods from simulation reults.