Vector quantiser of video signals

Abstract
An efficient design method is presented for a vector quantiser of 16-dimensional vectors based on a long training sequence of normalised video signals. The basic properties are demonstrated by examples.