Iterative algorithms for lossy source coding

Abstract

This thesis explores the problems of lossy source coding and information embedding. For lossy source coding, we analyze low density parity check (LDPC) codes and low density generator matrix (LDGM) codes for quantization under a Hamming distortion. We prove that LDPC codes can achieve the rate-distortion function. We also show that the variable node degree of any LDGM code must become unbounded for these codes to come arbitrarily close to the rate-distortion bound. For information embedding, we introduce the double-erasure information embedding channel model. We develop capacity-achieving codes for the double-erasure channel model. Furthermore, we show that our codes can be efficiently encoded and decoded using belief propagation techniques. We also discuss a generalization of the double-erasure model which shows that the double-erasure model is closely related to other models considered in the literature.