For many multimedia applications including VOD and image database, Fractal image compression algorithm has excellent characteristics in terms of high compression ratio and fast decompression speed but has been widely studied only in regard to software. However, because codec algorithm implemented by software is not able to reconstruct high resolution images in real-time, I suggest that design of dedicated hardware is necessary for fast decompression of moving images. In addition, Low power design is hot issue since portable devices need long-time operation with limited battery source. Therefore, I designed a low power quadtree fractal image decoder which can decode 256x256, 8 bits gray-scale images in real-time. Two power-down techniques have been adopted in the implementation, a power-down multiplier used in data-path, and a simple post-processing method optimized in terms of hardware utilization and energy dissipation. The decoder designed by using 1 poly, 3 metal, 0.6um CMOS technology can reconstruct 70~80 frames of 256x256, 8 bits gray-scale image in 1 sec with 3.3V, 50MHz.
