Invisible Modification of the Palette Color Image Enhancing Lossless Compression

Our contribution relates to lossless compression of pseudo color images (images with a palette). The proposed method is a preprocessing step preceeding actual compression. Indices in the palette are semioptimally permuted during preprocessing. For actual image compression, our own nonlinear predictor based method is used 1 but the proposed invisible palette modiication is relevant to most of other compression techniques too. Experiments with numerous images show that indices reordering in the palette yields data savings from 10 to 50 % for typical images. We suggest a preprocessing phase that (a) analyses statistics of the adjacency relations of index values, (b) performs optimization, and (c) permutes indices to palette to achieve more smooth image. The smoother image causes that the lossless image compression methods yield less output data. The task to optimally permute indices is a NP complete combinatorial optimization. Instead of checking all possibilities, we propose a reasonable initial guess and a fast suboptimal hill climbing optimization. 1. TASK FORMULATION AND PRINCIPAL IDEA 1.1. Palette as source of savings This paper discusses a lossless compression of pseudo color images (images with a palette). Many compression techniques utilize the assumption about the continuity of the image function. The same is true for our nonlinear predictor-based image compression 1 that uses image representation that is still not widely known. 2 Typical pseudo color images are reindexed by the palette and break continuity (or smoothness) assumption. Let us rst introduce a few concepts. The pseudocolor image is created by a mapping R; G; B] = palette(f(x; y)). R; G; B] are individual color components of the output pseudocolor image, i.e. three intensity images. The palette image consists of (1) the input index image f (x; y) and (2) a look up table called a palette. The f (x; y) will be called a palette index function in the sequel. The pseudocolor image is usually quantized from the true color image or is produced by an interactive painting program. The quantization algorithm causes discontinuities in the palette index function f (x; y). This eeect can signiicantly decrease the compression ratio. 1.2. Works of others There are many available compression algorithms for gray level images and true color images. Pseudo color images are usually compressed by the same algorithms as gray level ones. The compression ratio depends in these cases on a rst order entropy (we will call it smoothness in the sequel). A simple …