L. Anson. Fractal image compression. BYTE Magazine, 1993.
For several years, many people in the personal computing world have been chasing the Holy Grail of mass-market computing: multimedia. Two obstacles to reaching that goal have been overcome: 16- and 24-bit color printers and scanners are now readily available, and true-color graphics adapters display stunning pictures. But the large size of the image files required for these beautiful images is a huge problem. A single 800- by 600-pixel true-color image requires 1.44 MB of disk space; an uncompressed 10-second video clip with 30 frames per second at 320 by 200 pixels in true color requires an enormous 57.6 MB of disk space. Clearly, compression is necessary. JPEG is one commonly used compression technique. But you may not be aware of fractal image compression, which combines high compression ratios with fast decompression times. Compression technologies can be divided into lossless and lossy methods. A lossless method always produces a decompressed image that is identical, pixel-for-pixel, to the original image. The problem with lossless methods, such as the one used in PK Ware's PKZip, is that the attainable compression ratios on images are very small-typically 2 to 1. Lossy compression methods designed for image data can achieve much higher compression ratios. Both the DCT (Discrete Cosine Transform) and fractal-transform image-compression methods are lossy, but in other respects they are very different.
@article{Anso93,
Author = {Anson, L.},
Title = {Fractal image compression},
Journal = {BYTE Magazine},
Month = {},
Year = {1993}
}