The Variational Model Of Image Compression/decompression And Its Split Bregman Algorithm

Image compression/decompression is a classic problem of the image processing. Image inpainting restores image for missing information by retaining information, while image decompression restores the information by using non-redundant useful information. After analyzing the advantages of these two processes in detail, we can accomplish image decompressions combining with image inpainting. During decompressions, the compressed data can be seen as a damage image with part of information lost, and it will effectively decompressed from compressed images using image restoration techniques.Inspired by recent advancements in image inpainting techniques, we propose image compression models based on the above method. In this paper we propose new models with variation method and Split Bregman method to do a more in-depth study for image decompression technique. And the main achievements are as follows:Firstly, propose three new models for the image decompression by using different sampling methods, separately based on the first derivative, the second derivative, and combining these two which improves the image quality efficiently. Secondly, bring new algorithm to these three proposed models, with designing the corresponding fast Split Bregman method. Thirdly, a large number of numerical experiments in this paper have proved that the models and algorithm for decompression of this article is effective and feasible.