Analysis/synthesis techniques for subband image coding
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
The investigation focuses on the design of analysis/synthesis systems for image coding and the perceptual impact of these systems at low bit rates, and the theory, design, and implementation of both recursive and nonrecursive filtering systems are discussed.
Abstract
Analysis/synthesis systems designed for low bit rate image coding, their impact on overall system quality, and their computational complexity are discussed. The investigation focuses on the design of analysis/synthesis systems for image coding and the perceptual impact of these systems at low bit rates. Two objectives are emphasized in developing these systems: confining the total size of the subband images to be equal to the original image size, and designing the filters so that perceptual distortion is not introduced by the analysis/synthesis system. Methods based on circular convolution and symmetric extensions are developed and discussed in detail. The theory, design, and implementation of both recursive and nonrecursive filtering systems are discussed. Methods are introduced which display advantages over conventional quadrature mirror filter based approaches.>
