A study of DCT image coding using adaptive three‐dimensional scanning

Abstract

In video bandwidth compression, motion compensation and two-dimensional DCT have been used in a hybrid scheme. In this research, in place of the motion compensation of previous methods, a three-dimensional DCT is investigated that uses a DCT transform to reduce the interframe redundancy. When a three-dimensional DCT transform method is used for transmitting the DCT coefficients of video, the transmitted code rate and decoder signal-to-noise ratio can vary depending upon the three-dimensional scanning method used for the DCT coefficients. These parameters are influenced by both image detail and motion. An adaptive scanning method is proposed in which the three-dimensional scanning method is adaptively switched among four options. Simulation results show that compared to a fixed scanning method, in the adaptive approach, when all the postquantized coeffcients are transmitted (fixed image quality), the transmitted coded rate is economized, and even for the case in which the number of DCT coefficients that is transmitted is restricted, the reduction in image quality can be made very small.