Parallel implementation of fractal image compression

Digital images require a large amount of data to be represented. To make image storage and transmission practical and economical, image compression has become a major issue. In the past few years, several image compression methods using fractal theory have been developed. These methods promise better performances for compression. One of the most efficient approaches is based on iterated function systems (IFS), and has been promoted by Barnsley (1988). The basic idea is that an image can be reconstructed using the self-similarity it contains. A way to speed up the encoding process is to implement parallel processing using PVM (parallel virtual machine) software. The system utilizes both static and dynamic load allocations to obtain substantial compression time speedup over the original, sequential encoding implementation. In this paper, considerations such as PSNR, compression ratio, compression time versus number of processors and the workload granularity are also presented.