Abstract:This work is devoted to practical joint source channel coding. Although the proposed approach has more general scope, for the sake of clarity we focus on a specific application example, namely, the transmission of digital images over noisy binary-input output-symmetric channels. The basic building blocks of most state-of the art source coders are: 1) a linear transformation; 2) scalar quantization of the transform coefficients; 3) probability modeling of the sequence of quantization indices; 4) an entropy coding stage. We identify the weakness of the conventional separated source-channel coding approach in the catastrophic behavior of the entropy coding stage. Hence, we replace this stage with linear coding, that maps directly the sequence of redundant quantizer output symbols into a channel codeword. We show that this approach does not entail any loss of optimality in the asymptotic regime of large block length. However, in the practical regime of finite block length and low decoding complexity our approach yields very significant improvements. Furthermore, our scheme allows to retain the transform, quantization and probability modeling of current state-of the art source coders, that are carefully matched to the features of specific classes of sources. In our working example, we make use of ``bit-planes'' and ``contexts'' model defined by the JPEG2000 standard and we re-interpret the underlying probability model as a sequence of conditionally Markov sources. The Markov structure allows to derive a simple successive coding and decoding scheme, where the latter is based on iterative Belief Propagation. We provide a construction example of the proposed scheme based on punctured Turbo Codes and we demonstrate the gain over a conventional separated scheme by running extensive numerical experiments on test images.

A Unified Perspective on Parity- and Syndrome-Based Binary Data Compression Using Off-the-Shelf Turbo Codecs

A New Stopping Criterion for Duo-binary Turbo Codes

Quantum State Compression with Polar Codes

Robust Distributed Compression with Learned Heegard-Berger Scheme

BER comparison between Convolutional, Turbo, LDPC, and Polar codes

Breaking the Computational Bottleneck: Design of Near-Optimal High-Memory Spatially-Coupled Codes

Information-Theoretic Limits on Compression of Semantic Information

Zero-Error Distributed Function Compression for Binary Arithmetic Sum

Lossless text compression by means of binary-coded ternary number representation

Reconstruction of Turbo-code Interleaver at High Bit Error Rate

On the proper reconstruction of complex dynamical systems spoilt by strong measurement noise

High-Throughput Polar Code Decoders with Information Bottleneck Quantization

A Practical Approach to Lossy Joint Source-Channel Coding

Turbo and Turbo-Like Codes: Principles and Applications in Telecommunications

A Compressed Sensing-Based Low-Density Parity-Check Real-Number Code

An Iterative Decoding Scheme With Turbo Code And Iteratively Demapped Multi-Dimensional Qpsk Serially Concatenated

High Throughput Polar Code Decoders with Information Bottleneck Quantization

Bitstream Organization for Parallel Entropy Coding on Neural Network-based Video Codecs

Semantic Compression with Side Information: A Rate-Distortion Perspective

Parallel Data Compression Techniques