Higher order data compression with neural networks

A data compression method using a neural-network-based architecture is presented. This network will accept a large amount of scene or image data, compress it for storage or transmission, retrieve it and reproduce the original image. The architecture of the network is a feedforward four-layer network. The network is trained by using a nested training algorithm. The authors previously (1990) presented limited results on the performance and the operation of such a system. In this paper, the authors present an extensive detailed performance evaluation by computer simulation of this system using 512×512 pixels and 8 bits/pixel images