Overlap and channel errors in Adaptive Vector Quantization for image coding

Abstract

Adaptive vector quantization (AVQ) is a flexible adaptive single pass image compression method, with fast decompression time, for which practical compression achieved compares favorably with existing methods such as the JPEG standard and traditional trained VQ. In this paper we analyze its asymptotic performance for the lossless case and present variations to limit the effect of transmission errors on the decoding image.