Nowadays, more and more images from every area need to be stored. If a lossless compression of the image is desired, the storage space requirements can still be enormous. The aim of this paper is to show how to store those images in such a way that a quick access AND a big storage capacity are available together. The tools used are a multi-resolution entropy coding (the Haar transform and the Universal Variable Length Coder) combined with a hierarchical quantization (the slicing) of each resolution. The main idea is to have a small icon stored on a fast access media, instantly accessible, while the high resolution components are stored on mass media and are loaded only if needed. The total compression is lossless.