COUNTER-PROPAGATION NEURAL-NETWORK FOR IMAGE COMPRESSION

Recently, several image compression techniques based on neural network algorithms have been developed. In this paper, we propose a new metho d for image compression-the modified counter-propagation neural networ k algorithm, which is a combination of the self-organizing map of Koho nen and the outstar structure of Grossberg. This algorithm has been su ccessfully used in many applications. The modification presented has a lso demonstrated an interesting performance in comparison with the sta ndard techniques. It was found that al the learning stage we can use a ny image for a network training (without a significant influence on th e net operation) and the compression ratio and quality depend on the s ize of the basic element (the number of pixels in the cluster) and the amount of error tolerated when processing. (C) 1996 Society of Photo- Optical Instrumentation Engineers.