Vector quantization based on visual masking functions for the DPCM and HDPCM coding of coloured pictures

In DPCM (differential pulse code modulation) or HDPCM (hybrid DPCM) of coloured pictures consisting of three components, every component can be coded with three separate coders, each containing a scalar quantizer. However, more compression can be obtained with one vector quantizer instead of three scalar quantizers. In the article, the attention is focused on the placement of the output vectors in the two-dimensional prediction error plane using visual masking functions.<
>