Block classification method based on human perception for dynamic bitstream scaling of MPEG video

Abstract

This paper proposes a classification method of blocks in an image for the purpose of dynamic spatial bitstream scaling of MPEG video based on simple characteristics of human perception. First, we classify the blocks based on the relation between the characteristics of distorted blocks difficult to be perceived by human and the characteristics of blocks derived by the moving vectors. Then, based on this relation, we propose the bitstream scaling method for MPEG video by stripping off the DCT coefficients. An experiment using this techniques in an actual media synchronization system has illustrated the usefulness of the proposed approach.