In this paper a neural network based image compression method is presented. Neural networks offer the potential for providing a novel solution to the problem of data compression by its ability to generate an internal data representation. Our network, which is an application of counter propagation network, accepts a large amount of image data, compresses it for storage or transmission, and subsequently restores it when desired. A new approach for reducing training time by reconstructing representative vectors has also been proposed. Performance of the network has been evaluated using some standard real world images. It is shown that the development architecture and training algorithm provide high compression ratio and low distortion while maintaining the ability to generalize and is very robust as well. 1

Quantitative neural networks are derived from psychological postulates about punishment and avoidance. The classical notion that drive reduction is reinforcing is replaced by a precise physiological altetAative akin to Miller's "Go " mechanism and Estes's "amplifier " elements. Cell clusters d} and dj are introduced which supply negative and positive incentive motivation, respectively, for classical conditioning of sensory-motor acts. The dj cells are persistently turned on by shock (on-cells). The dj cells a,re transiently turned on by shock termination (off-cells). The rebound from d} cell activation to d1 cell activation replaces drive reduction in the case of shock. Classical conditioning from sensory cells.,9? to the pattern of activity playing on arousal cells d I = (dj, dj) can occur. Sufficiently positive net feedback from d I to.,9? can release sampling, and subsequent learning, by prescribe.d cells in.,9? of motor output controls. Once sampled, these controls can be reactivated by.,9? on recall trials. This concept avoids some difficulties of two-factor theories of punishment and avoidance. Recent psychophysiological data and concepts are analyzed in terms of network