Block sizes of practical vector quantization (VQ) image coders are not larg
e enough to exploit all high-order statistical dependencies among pixels. T
herefore, adaptive entropy coding of VQ indexes via statistical context mod
eling can significantly reduce the bit rate of VQ coders for given distorti
on, Address VQ was a pioneer work in this direction. In this paper we devel
op a framework of conditional entropy coding of VQ indexes (CECOVI) based o
n a simple Bayesian-type method of estimating probabilities conditioned on
causal contexts. CECOVI is conceptually cleaner and algorithmically more ef
ficient than address VQ, with address-VQ technique being its special case,
It reduces the bit rate of address VQ by more than 20% for the same distort
ion, and does so at only a tiny fraction of address VQ's computational cost
.