An algorithm for noiseless image coding with application to a smooth approximation of 3-dimensional surfaces

Image data could be modeled by a collection of patches. The proposed universal data compression algorithm takes advantage of this spatial structure of images, and employs a two stage minimum description length principle. The coding scheme enables two stage progressive transmission of images. Furthermore if we plot an image by pixel positions in the x-y plane and its intensity in the z axis, it is a complex surface in 3-dimensional space. Our modeling gives a smooth approximation of this surface and the algorithm could be applied to varieties of surface smoothing problems. The computational complexity of the algorithm is O(Mlog M), where M is the number of pixels