arXiv subjects

Lorenzo Cappellari

Publications and source records attributed to Lorenzo Cappellari.

6 recordsLinked to original sources

arXiv2010

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

We consider the problem of compressing memoryless binary data with or without side information at the decoder. We review the parity- and the syndrome-based approaches and discuss their theoretical limits, assuming that there exists a virtual binary symmetric channel between the source and the side information, and that the source is not necessarily uniformly distributed. We take a factor-graph-based approach in order to devise how to take full advantage of the ready-available iterative decoding procedures when turbo codes are employed, in both a parity- or a syndrome-based fashion. We end up obtaining a unified decoder formulation that holds both for error-free and for error-prone encoder-to-decoder transmission over generic channels. To support the theoretical results, the different compression systems analyzed in the paper are also experimentally tested. They are compared against several different approaches proposed in literature and shown to be competitive in a variety of cases.

cs.IT↗

arXiv2010

Lossy Source Compression of Non-Uniform Binary Sources Using GQ-LDGM Codes

In this paper, we study the use of GF(q)-quantized LDGM codes for binary source coding. By employing quantization, it is possible to obtain binary codewords with a non-uniform distribution. The obtained statistics is hence suitable for optimal, direct quantization of non-uniform Bernoulli sources. We employ a message-passing algorithm combined with a decimation procedure in order to perform compression. The experimental results based on GF(q)-LDGM codes with regular degree distributions yield performances quite close to the theoretical rate-distortion bounds.

cs.IT↗

arXiv2010

On Syndrome Decoding for Slepian-Wolf Coding Based on Convolutional and Turbo Codes

In source coding, either with or without side information at the decoder, the ultimate performance can be achieved by means of random binning. Structured binning into cosets of performing channel codes has been successfully employed in practical applications. In this letter it is formally shown that various convolutional- and turbo-syndrome decoding algorithms proposed in literature lead in fact to the same estimate. An equivalent implementation is also delineated by directly tackling syndrome decoding as a maximum a posteriori probability problem and solving it by means of iterative message-passing. This solution takes advantage of the exact same structures and algorithms used by the conventional channel decoder for the code according to which the syndrome is formed.

cs.IT↗

arXiv2009

On Superposition Coding for the Wyner-Ziv Problem

In problems of lossy source/noisy channel coding with side information, the theoretical bounds are achieved using "good" source/channel codes that can be partitioned into "good" channel/source codes. A scheme that achieves optimality in channel coding with side information at the encoder using independent channel and source codes was outlined in previous works. In practice, the original problem is transformed into a multiple-access problem in which the superposition of the two independent codes can be decoded using successive interference cancellation. Inspired by this work, we analyze the superposition approach for source coding with side information at the decoder. We present a random coding analysis that shows achievability of the Wyner-Ziv bound. Then, we discuss some issues related to the practical implementation of this method.

cs.IT↗

arXiv2007

Trellis-Coded Quantization Based on Maximum-Hamming-Distance Binary Codes

Most design approaches for trellis-coded quantization take advantage of the duality of trellis-coded quantization with trellis-coded modulation, and use the same empirically-found convolutional codes to label the trellis branches. This letter presents an alternative approach that instead takes advantage of maximum-Hamming-distance convolutional codes. The proposed source codes are shown to be competitive with the best in the literature for the same computational complexity.

cs.IT↗

arXiv2007

Distributed Source Coding Using Continuous-Valued Syndromes

This paper addresses the problem of coding a continuous random source correlated with another source which is only available at the decoder. The proposed approach is based on the extension of the channel coding concept of syndrome from the discrete into the continuous domain. If the correlation between the sources can be described by an additive Gaussian backward channel and capacity-achieving linear codes are employed, it is shown that the performance of the system is asymptotically close to the Wyner-Ziv bound. Even if such an additive channel is not Gaussian, the design procedure can fit the desired correlation and transmission rate. Experiments based on trellis-coded quantization show that the proposed system achieves a performance within 3-4 dB of the theoretical bound in the 0.5-3 bit/sample rate range for any Gaussian correlation, with a reasonable computational complexity.

cs.IT↗