A variable-rate image coding scheme with vector quantization and clustering interpolation

The authors propose a simple image coding algorithm using VQ (vector quantization) and clustering interpolation, which avoids an explicit interpolation operation by employing different but related codebooks at the encoder and the decoder of a vector quantizer. Computer simulations indicated that images reconstructed with clustering interpolation had smaller mean square errors than those obtained with linear or contour-based interpolation. In order to reduce coding artifacts and to improve reconstruction quality at low bit rates, an adaptive corrector stage with block classification is also included. Computer simulations demonstrate the effectiveness of the algorithm at about 0.3 b/pixel.<>