In this paper an algorithm for grayscale image compression based on usage of two optimal fixed quantizers and one uniform adaptive quantizer designed for discrete input samples is presented. Applying this algorithm to standard grayscale images, experimental results are obtained. The results show that reduction of the bit-rate is more than 50 %, which is a significant improvement compared to the previously published paper. Reconstructed images have very high level of quality i.e. near lossless compression is achieved.

Keywords: Grayscale images, Compression, Multimode coding, Forward adaptation