Image Compression based on WFA



Original 512x512, 24 bpp



Reconstructed image, 213x compressed, PSNR = 24.91 dB



Reconstructed image (high compression), compressed 642x, PSNR = 22.55 dB


The compression algorithm has been aquired by industry and will be not available for public.

Image Compression based on GFA and Vector Quantization



Left: Original 512x512, 1 bpp
Right: Reconstructed, 40x compressed, 836 wrong pixels.

Left: Original 512x512, 3 bpp
Right: Reconstructed, 154x compressed, 638 wrong pixels.


Note: my DCC'96 codec is now available. For a detailed description of the algorithm see Finite Automata Based Compression of Bi-level and Simple Color Images written by K. Culik II and V. Valenta.



This page is under construction.


Go to Vladimir Valenta's Home Page Go to Karel Culik's Home Page