A fast exact GLA based on code vector activity detection

This paper introduces a new method for reducing the number of distance calc ulations in the generalized Lloyd algorithm (GLA), which is a widely used m ethod to construct a codebook in vector quantization, Reduced comparison se arch detects the activity of the code vectors and utilizes it on the classi fication of the training vectors. For training vectors whose current code v ector has not been modified, we calculate distances only to the active code vectors. Large proportion of the distance calculations can be omitted with out sacrificing the optimality of the partition, The new method is included in several fast GLA variants reducing their running times over 50% on aver age.