Jim Chou

Duality between source coding and channel coding and its extension to the side information case

We explore the information-theoretic duality between source coding with side information at the decoder and channel coding with side information at the encoder. We begin with a mathematical characterization of the functional duality between classical source and channel coding, formulating the precise conditions under which the optimal encoder for one problem is functionally identical to the optimal decoder for the other problem. We then extend this functional duality to the case of coding with side information. By invoking this duality, we are able to generalize the result of Wyner and Ziv (1976) relating to no rate loss for source coding with side information from Gaussian to more arbitrary distributions. We consider several examples corresponding to both discrete- and continuous-valued cases to illustrate our formulation. For the Gaussian cases of coding with side information, we invoke geometric arguments to provide further insights into their duality. Our geometric treatment inspires the construction and dual use of practical coset codes for a large class of emerging applications for coding with side information, such as distributed sensor networks, watermarking, and information-hiding communication systems.

Turbo and trellis-based constructions for source coding with side information

The problem of rate-distortion efficient constructions is studied for the problem of source coding with side information (SCSI), which has assumed heightened interest. While the Wyner-Ziv theorem from information theory has prescribed rate-distortion performance bounds for the SCSI problem, the gap between theory and practice has remained large. To reduce this gap, two different frameworks are proposed based on a trellis construction and a turbo-based construction respectively. Simulation results on the Gaussian SCSI problem reveal the promise of the proposed approaches: at 1 bit per sample, 0.5 bits/sample, 0.25 bits/sample and 0.125 bits/sample, these constructions attain performance within 1.3 dB, 1.1 dB, 0.85 dB and 0.5 dB respectively of the theoretical Wyner-Ziv rate-distortion bound.

A robust blind watermarking scheme based on distributed source coding principles

We propose a powerful new solution to the multimedia watermarking problem by exploiting its duality with another problem for which we have recently made pioneering constructive contributions. This latter problem is that of distributed source coding, or compression of correlated sources that are distributed. We show how these two seemingly unrelated problems are actually duals of each other. We exploit this duality by transforming our recently introduced powerful constructive framework for the distributed compression problem in [13] to a corresponding dual framework for the watermarking problem. Simulations expose the significant performance gains attained by our proposed watermarking approach and reveal its exciting potential for next-generation watermarking techniques. This can be accredited to the exploitation of the dual roles played by source codes and channel codes in the two problems.

Tracking and exploiting correlations in dense sensor networks

In this paper, we propose a novel method for reducing energy consumption in a sensor network. It is important in a sensor network to minimize the energy usage of each sensor, because the nodes typically have finite battery life and if a node dies, this can lead to a loss of data or a network partition. As a result, several researchers have proposed various methods of routing and communication between nodes to reduce energy consumption. We propose an orthogonal approach to previous methods. In particular, we propose to exploit the inherent correlations that exist between sensor nodes by devising a novel algorithm that enables sensor nodes to compress their readings without knowing the exact measurements of the other nodes. Our simulations show that our algorithm used is promising as it leads to significant energy saving for various types of sensor nodes.

Turbo coded trellis-based constructions for data embedding: channel coding with side information

A host of emerging applications in multimedia and communications are connected to the problem of channel coding with side information (CCSI). In CCSI, side information about the transmission channel is present at the sender but not at the receiver. The goal of this paper is to bridge the gap between practical code constructions and the theoretical capacity bounds. Specifically, we present a new turbo-coded trellis-based code construction that appears to be very promising: for the Gaussian CCSI problem at a transmission rate of 1 bit/channel use, our proposed approach comes within 2.72 dB of the information-theoretic capacity established by M. Costa (see IEEE Trans. on Information Theory, vol.29, p.439-41, 1983).

A distributed and adaptive signal processing approach to exploiting correlation in sensor networks

Abstract We propose a novel approach to reducing energy consumption in sensor networks using a distributed adaptive signal processing framework and efficient algorithm. 1 While the topic of energy-aware routing to alleviate energy consumption in sensor networks has received attention recently [C. Toh, IEEE Commun. Mag. June (2001) 138; R. Shah, J. Rabaey, Proc. IEEE WCNC, March 2002], in this paper, we propose an orthogonal approach to complement previous methods. Specifically, we propose a distributed way of continuously exploiting existing correlations in sensor data based on adaptive signal processing and distributed source coding principles. Our approach enables sensor nodes to blindly compress their readings with respect to one another without the need for explicit and energy-expensive inter-sensor communication to effect this compression. Furthermore, the distributed algorithm used by each sensor node is extremely low in complexity and easy to implement (i.e., one modulo operation), while an adaptive filtering framework is used at the data gathering unit to continuously learn the relevant correlation structures in the sensor data. Applying the algorithm to testbed data resulted in energy savings of 10–65% for a multitude of sensor modalities.

Robust turbo-based data hiding for image and video sources

Channel coding with side information (CCSI) provides a powerful framework for constructing data embedding codes. As a result, there has been a considerable amount of work done in trying to apply these CCSI codes to the field of image and video watermarking. The CCSI codes that exist in the literature (see Chou, J. et al., 1999; Kesal, M. et al., 2000; Eggers, J. et al., 2001), however, are relatively far from theoretical bounds. Furthermore, these codes, by themselves, are not well-suited for dealing with geometrical distortions which often appear in watermarking attacks. We propose a CCSI code that is closer to the theoretical bounds (within 2.0 dB) than the codes that exist in the literature. In addition, we show how this code construction can be combined with a synchronization code to deal with geometrical distortions.

High capacity audio data hiding for noisy channels

We combine theoretical and algorithmic advances in the area of information-hiding with the current mature knowledge-base in the human audio perception system to propose a novel audio data hiding technique that significantly pushes the state-of-the-art in the field. Our work is based on a combination of advances in two disjoint fields: information hiding and human auditory masking. The field of information hiding has seen a resurgence due to advances in the understanding of fundamental bounds from information theory. By integrating this with the human perceptual system knowledge that has been successfully exploited for several years in the audio compression community, we derive a new and improved audio data hiding technique that finds application in a number of exciting scenarios like music enhancement and digital communications over analog data channels. Our preliminary results show that we can embed data at an order of magnitude higher rate than existing audio data hiding systems, while being robust to channel noise.

Next generation techniques for robust and imperceptible audio data hiding

We combine recent theoretical and algorithmic advances in the area of information-hiding with the current mature knowledge-base in the human audio perception system to propose a novel audio data-hiding technique that significantly pushes the state-of-the-art in the field. Our work is based on a combination of advances in two disjoint fields: information-hiding and human auditory masking. The field of information-hiding has recently seen a resurgence due to advances in the understanding of fundamental bounds from information theory. By integrating this with the human perceptual system knowledge that has been successfully exploited for several years in the audio compression community, we derive a new and improved audio data-hiding technique that finds application in a number of exciting scenarios like music enhancement and digital communications over analog data channels. Our preliminary results show that we can embed data at a rate an order of magnitude higher than existing audio data hiding systems, while being robust to channel noise.

Watermarking based on duality with distributed source coding and robust optimization principles

Inspired by a previously proposed constructive framework for the distributed source coding problem. We propose a powerful constructive approach to the watermarking problem, emphasizing the dual rules of distributed source coding with side information at the decoder and channel coding with side information at the encoder. In our framework, we explore various source and channel codes to close the gap on the achievable capacity of watermarking systems. We propose two methods of solution, one which is based on optimal rate-distortion quantizers and the other based on robust optimization and convex programming. The resulting watermarking schemes, when subjected to additive white gaussian noise (AWGN) attacks, achieve results which are comparable to or better than the best watermarking schemes in the literature.