Localized model-based approach to teleconferencing using three-dimensional vector quantization

Unlike the case of speech, the use of models in image processing and video coding is rather limited, mainly due to the inherent difficulty in characterizing image sources. A new model-based approach to video teleconferencing is introduced using three-dimensional vector quantization (VQ). According to the proposed model, the recent localized history of the sequences is used to encode the present features based on a hierarchical approach. The results indicate that a high compression ratio of over 100:1 with high quality of the perceived sequence is possible.