Tutorials

Tutorial title: Distributed Video Coding: Basics, Codecs and Applications

Authors: Christine Guillemot and Aline Roumy

Affiliation: AINRIA/IRISA, Campus Universitaire de Beaulieu, Rennes, France.

Tutorial outline:

Distributed source coding has emerged as an enabling technology for sensor networks. It refers to the compression of correlated signals captured by different sensors which do not communicate between themselves. All the signals captured are compressed independently and transmitted to a central base station which has the capability to decode them jointly. Distributed source coding finds its foundation in the seminal work of Slepian-Wolf (1973) and Wyner-Ziv (1976). The minimum achievable rate for lossless compression of two statistically dependent memoryless sources is given by the joint entropy of the two sources. Slepian and Wolf have established that this lossless compression rate bound could be approached with a vanishing error probability for long sequences, even if the two sources are coded separately, provided that they are decoded jointly and that their correlation is known to both the encoder and the decoder. This theorem has led to a new coding paradigm known as distributed source coding. The lossy equivalent of the Slepian-Wolf theorem has been formulated later on by Wyner and Ziv.

The proof of the Slepian-Wolf W theorem is based on random binning, which is non-constructive, i.e., it does not reveal how practical code design should be done. In 1974, Wyner suggested the use of parity check codes to approach the corner points of the Slepian-Wolf rate region. It is only recently that practical solutions based on channel capacity-achieving codes (block codes, turbo codes or LDPC codes) have been explored for applications ranging from video compression, resilient video transmission, to minimization of transmit energy in sensor networks.

Video compression, as well as scalable video compression, has been recast into a distributed source coding framework leading to distributed video coding schemes targeting mainly low coding complexity and error resilience functionalities. Correlated samples (pixels or transform coefficients) from different frames are regarded as outputs of different sensors. However, the application of the Wyner-Ziv principles to video compression is not straightforward and requires solving a number of issues. The Distributed Source Coding principles apply quite naturally to the compression of video sequences captured of the same scene by several cameras. With respect to classical multiview coding techniques, DVC allows the exploitation of correlation between views without - or with limited - inter-sensor (that is inter-camera) communication.

This tutorial will present the underlying theory as well as latest developments of distributed video compression for both monoview and multiview applications.

Information about the authors:

Christine Guillemot is currently 'Directeur de Recherche' at INRIA, in charge of the TEMICS research group dealing with image modelling, processing, video communication and watermarking. She holds a PhD degree from ENST (Ecole Nationale Superieure des Telecommunications) Paris. From 1985 to October 1997, she has been with FRANCE TELECOM/CNET, where she has been involved in various projects in the domain of coding for TV, HDTV and multimedia applications, and co-ordinated a few (e.g. the European RACE-HAMLET project). From January 1990 to mid 1991, she has worked at Bellcore, NJ, USA, as a visiting scientist. Her research interests are signal and image processing, video coding, and joint source and channel coding for video transmission over the Internet and over wireless networks. She has served as Associate Editor for IEEE Trans. on Image Processing (2000-2003), and for IEEE Trans. on Circuits and Systems for Video Technology (2004-2006). She is a member of the IEEE IMDSP and of the IEEE MMSP technical committees.

Aline Roumy received the Engineering degree from Ecole Nationale Supérieure de l'Electronique et de ses Applications (ENSEA), Cergy, France in 1996, the Master degree in June 1997 and the Ph.D. degree in September 2000 from the University of Cergy-Pontoise, France. During 2000-2001, she has been the recipient of a French Defense DGA/DRET postdoctoral fellowship and was a research associate at Princeton University, Princeton, NJ. On November 2001, she joined INRIA, Rennes, France. Her current research and study interests include the area of statistical signal processing, coding theory and information theory.