Glicbawls - grey level image compression by adaptive weighted least squares

Glicbawls is an algorithm for losslessly and near-loesslessly compressing photographic greyscale images. It is based on locally adaptive least squares predictors, based on a two-dimensional exponentially fading memory. Despite being of moderate computational complexity, Glicbawls achieves excellent compression performance (e.g. the lenna image at 3.90 bits/pixel). The Glicbawls algorithm has been implemented in 1795 bytes of C code, making it feasible to include the decoder with the compressed images, and thus to determine a firm upper bound for the information content of any given image.