Sample selection in subband vector quantization
M. H. Ahmad Fadzil, Tim J. Dennis · International Conference on Acoustics, Speech, and Signal Processing · 2002
A sample selection procedure for a subband vector-quantization (VQ) interframe predictive coder incorporating short-kernel subband filter banks is presented. The coder is based on two-dimensional subband decomposition of the error image obtained after motion-compensated interframe prediction (MCIP). The luminance and chrominance subbands, generated after the decomposition process, form image vectors which are then quantized using conventional vector quantization. By using the sample selection procedure, selective quantization of the image vectors can be carried out with minimal distortion introduced into the reconstructed images. Any major spatial redundancies in the MCIP error image are exploited to achieve the low output bit rate desired.>