Fractal Image Compression Using a Circulating Pipeline Computation Model

In this paper, we present a model, and experimental results, for performing fractal image compression using a circulating pipeline computation model. For this model, a toroidal linear array of processors is employed and utilized in a pipelined fashion. Image data decomposition is performed as to minimize memory requirements for a given processor in the pipeline and to obtain reasonable load balance throughout the pipeline. In addition to parallel decomposition, a simple image classification scheme is employed to reduce computational complexity. Quantitative results include an evaluation of attained compression ratios, signal-to-noise ratios (SNR) for reconstructed images, comparison with JPEG compression, and performance of the parallel system with respect to achieved execution times and efficiency. Experimental results, using an nCUBE-2 supercomputer, are presented.