On successively refinable trellis-coded quantization

There are many applications in which scalable encoding and progressive transmission of information are very attractive. Successively refinable quantization, as a strategy for scalable coding, seeks for an embedded optimal representation of data in the rate-distortion sense. In this paper, successively refinable trellis-coded quantization (TCQ) is studied. Motivated by the observations on the TCQ quantization residual, and through characterizing the geometry of TCQ Voronoi regions and exploiting the algebraic structure of trellis, an algorithm is proposed, which is able to achieve successive refinability for TCQ at high quantization rates. Simulation results show that the proposed algorithm has excellent performance.