Self-organization feature map based on VQ components to solve image coding problem

Abstract

This paper present image coding which is gained many researchers attention in order to improve the quality of image after the compression process. Since this is expended most computing resources and research which is related not only to search for a mathematical transformation, but also to study the characteristics of visual perception of the image features and fail-safe transmission of images via communication channels. There are many methods of image coding with neural networks of 2D SOFM kohonen map have been suggested and investigated. The coding schemes are proposed methods vector quantization as the original image, and the spatial frequency image component derived from the adaptive to the contours of the two-dimensional analysis and synthesis. The calculation of the computational cost in compression based on Kohonen maps. The methods are characterized by a high level of adaptation due to the introduction of educational stage that provides for the increase of multiplication ratio and high quality of image restarting after coding. The modified method of image multiplexing based on characteristic feature of the given method is vector digitizing of image components. This paper considers the coding problem of photo realistic images, presented in a digital form. The characteristic feature of the method is the application of pair exchange, this increases processing speed and sorting of data arrays. However the result of proposed method is shown the image quality after compression processor. Using this approach the differences or lost pixels between the image after and before compression processer are considered. The propose method may useful for image representation and image coding researcher and such related field.