High fidelity and lower bit-rate predicted image coding system

Vector quantization (VQ) is a popular method commonly used in low cost compression of still image. However, there are a number of disadvantages such as high complexity of the encoder and the blocking effect in the decoded image. These problems can be partially overcome by predicted vector quantization (PVQ) which employs spatial prediction to reduce the complexity of the encoder and to improve the pictorial quality. Three novel predictive schemes that are applied to the existing PVQ system are introduced. From the experimental results, the improvement of the suggested scheme is significant both in quantitative and qualitative tests.