tection localization and tracking, including Global Posi- ... tions and Monitoring Systems ... fore there is nothing new other than yet another buzz-word. Of course ...
884
IEEE SIGNAL PROCESSING LETTERS, VOL. 19, NO. 12, DECEMBER 2012
Rolling Out the New EDICS
D
EAR reader, our journal is finally catching up with time, and after 6 years its EDICS are totally renewed. It is the result of months of consultation with the Senior Area Editors and Associate Editors and of the hard work of the IEEE Signal Processing Staff, that has allowed a much smooth transition than I ever expected. I want to express my sincere thanks to everybody involved for their help. I also invited the Senior Area Editors to submit editorial articles to highlight the areas that are of particular interest to them in these EDICS. You will read three of them on this issue and I hope to include more later on during the next year. I want to start with an area where the change was, perhaps, most needed. I am talking about Biomedical signal processing (BIO), whose EDICS underwent a total transformation, under the leadership of Jean-Christophe Olivo-Marin. This area now includes the following EDICS: • BIO-BIOIM Biological imaging • BIO-GENO Computational imaging • BIO-COMPIM Genomics/proteomics signal processing • BIO-INFO Bioinformatics • BIO-MAID Signal processing methods for medical aids • BIO-MDIG Medical diagnostic methods • BIO-SENS Sensor systems for biomedical signal and image processing • BIO-SSIGP Sensor signal processing • BIO-SBM Statistical biological modeling • BIO-BCI Signal Processing Methods supporting Brain/Human-Computer Interfaces There are several systematic problems in attracting more high quality papers in this field, but this is a small and long overdue step in the right direction. The Statistical Signal Processing Area has expanded its name and became the Machine Learning and Statistical Signal Processing Area, MLSAS for short, and it now incorporate five new EDICS: • MLSAS-SPARSE Theory of sparse representations and compressed-sensing • MLSAS-BAYES Bayesian learning; Particle filters; Information-theoretic learning • MLSAS-GRAPH Graphical and kernel methods • MLSAS-MAN Manifold learning, applications of Differential Geometry and adaptive signal processing on Manifolds • MLSAS-SOC Social networks, Social Learning models and Game Theoretic analysis • MLSAS-PATT Pattern recognition and classification Many thanks to Mark Coates, Kenneth Kreutz-Delgado, Konstantinos I. Diamantaras, and Michael Elad, for their great suggestions. Particularly dear to me are MLSAS-MAN and MLSAS-SOC, which well exemplify two types of objectives we had in renewing the list. One was adding mathematical
depth and breath to unleash new signal processing problems formulations. The other was adding applications of potentially high impact. As social networks are increasingly becoming dependent on information networks to mold their behavior, signal processing technology can more directly be tailored to shape the future landscape of social interactions. The Digital and multirate signal processing area, is still retaining the classic denomination DSP, but adding three much needed new topics: • DSP-GRAPH Graph analysis, Spectral Graph Theory and Algebraic Topology algorithms, Wavelets over graphs • DSP-FRI Finite rate of innovation sampling and processing of signals • DSP-SPARSE Sparse signal representations and recovery—algorithms and applications I invite you to read the editorials by Michael Elad and Thomas Stromer that follow, providing an extensive overview of the state of the art in compressive sensing and finite rate of innovation sampling, which the previous EDICS of MLSAS-SPARSE and the two EDICS, DSP-SPARSE and DSP-FRI, and other EDICS which I will mention later, are linked to. These are perhaps the hottest keywords in signal processing research, and our strategy to capture them was to divide and conquer. EDICS mentioning sparsity are spread in the different areas that these concepts have significantly contributed to advance. On a different note, my personal favorite in DSP is DSPGRAPH. I view graph structures as the new support of interest for discrete multidimensional signals; graphs can describe a rich set of relationships other than temporal and spatial proximity, and can also be a proxy for an underlying non linear relationship among data. I really wish to see more papers emerging in this area. One of the important connections signal processing thrives on is the close relationship to the physical world and with the hardware world. Three new EDICS have been added in the area of Signal Processing Design and Implementation area (HDW), with help from Tokunbo Ogunfunmi: • HDW-SECU Hardware security • HDW-PROG Programmable Hardware (e.g., FPGA, SoC, ASIC, etc.) for DSP algorithms • HDW-MCORE Multicore Processors for DSP algorithms In this set, Patrizio Campisi contributed the EDICS HWD-SECU. More importantly, thanks to Patrizio, the area of security in general has significantly expanded its reach in the new EDICS, with a new EDICS in Communications (COM) covering physical layer security and many added in the list of new EDICS Multimedia Signal Processing (MSP), where all the new EDICS except the one on Network Utility Maximization, are directed at representing better advances in the field of data security: • MSP-WAT Watermarking (Theory, Algorithms, Attacks) • MSP-STE Steganography and steganalysis
Digital Object Identifier 10.1109/LSP.2012.2224517 1070-9908/$31.00 © 2012 IEEE
IEEE SIGNAL PROCESSING LETTERS, VOL. 19, NO. 12, DECEMBER 2012
• MSP-HAS Hashing for content authentication/identification • MSP-NUM Network Utility Maximization algorithms, Resource Allocation and Game Theory for Media Content distribution systems • MSP-FOR Multimedia Forensics • MSP-SECU Security and Privacy preserving signal processing Patrizio also suggested the important EDICS on Signal and Image & Video Biometric Analysis (IMD-ARS-BIM) and on Image and Video Forensic Analysis, in the Image and Multidimensional Signal Processing area (IMD-ARS-FOR). Hayder Rhada was also instrumental in updating both the MSP as well as the IMD edics where, of course, we also added a new one on sparsity: • IMD-SPAR Sparse representation in Imaging • IMD-ARS-BIM Signal, Image & Video Biometric Analysis • IMD-ARS-SRE Image & Video Storage and Retrieval • IMD-ARS-SRV Image & Video Synthesis, Rendering, and Visualization • IMD-ARS-FOR Image & Video Forensic Analysis I am extremely grateful to Dilek Hakkani-Tür, in the Speech and Language Processing areas (SPE), and to Rudolf Rabenstein, in the Audio and Acoustic Signal Processing areas (AEA), to have brought new lymph to the EDICS in two of the absolutely core areas of Signal Processing, that have continued to shape the fame and fortune of our society for many years. Dilek suggested to add the following EDICS: • SPE-SLID Speaker/Language identification and diarization • SPE-SPL Spoken language processing The Audio area EDICS are completely revised: • AEA-RES Audio and Speech signals restoration • AEA-FEC Feedback and Echo Cancellation • AEA-AUD Auditory Models and Anthropomorphic Processing • AEA-SEP Source Separation and Enhancement • AEA-COD Audio Coding and Transmission • AEA-MIR Content-based Processing and Music Information Retrieval • AEA-AMS Audio Analysis, Modification, and Synthesis • AEA-MUL Multichannel Audio Processing • AEA-SAR Spatial Audio Recording and Reproduction • AEA-AAE Analysis of Acoustic Environments • AEA-SPARSE Audio processing via sparse representations Rudolf, together with Mads Græsbøll Christensen have written a wonderful editorial that gives the pulse of what is audio signal processing is in the new century and I am looking forward to expanding the pool of authors and readers in all these areas. Classic areas of Signal Processing are also Sensor array and multichannel signal processing (SAM) and Communications (COM). I am grateful to Magnus Jansson, Eduard Jorswieck, Vincent Lau and Peter Willet for their careful oversight on these changes, many of which I was personally invested in, given that these are the areas I am closer to. In the Signal Processing are also Sensor array area new EDICS are
885
• SAM-NET Distributed processing and optimization over sensor networks, gossiping algorithms • SAM-BEAM Beam-forming, including Space-time adaptive methods • SAM-DOAE Direction of arrival estimation, source detection localization and tracking, including Global Positioning Systems • SAM-LOC Sensor network localization algorithms • SAM-RASO Radar and Sonar Signal Processing • SAM-IMG Imaging with array data including Synthetic Aperture Radar • SAM-SPARCS Compressed Sensing A part from some revisions, in Communications we added: • COM-GREEN Energy Aware Communications and Networking • COM-SECU Communication Security and Privacy in Physical Layer communication and Networks • COM-NDIST Distributed coding for networks and databases • COM-RELN Communications over Relay Channels, Network coding • COM-NETA Analysis of Networks and Queuing systems • COM-OPTNE Optimal network resource allocation, Network Utility Maximization and Game theoretic models • COM-MACOG Multiple Access, Scheduling, Cognitive Radios These additions reflect an established trend towards looking at signal processing tools as having broader impact on several layers of the network architecture. I want to bring your attention to COM-GREEN and to other two EDICS that I first hesitated, but then decided to include. They are very important to me, but some people feel that they are too trendy (a more flattering version of flaky) and hearing this too many times I was tempted to, as they say, “take my marbles and go home”. But thanks to the solicitation of my Associate Editors, I resisted that temptation, and decided to add COM-GREEN and a new section on Emerging Signal Processing Applications (ESP) with the following two EDICS: • ESP-CPS Cyber-Physical Systems, Machine Communications and Monitoring Systems • ESP-EN Energy and Smart Grid My view is that we are at a critical time to modernize the old stodgy term of cybernetics and that we, with our good friends in Information Theory and Automatic Control, have the tools, if not the will, to break the mold. Unfortunately, at times, having a great legacy is what stands in the way. The objection that rings in my hears is that energy is what we always optimized, and therefore there is nothing new other than yet another buzz-word. Of course, looking at the very limits of performance of information systems versus signal to noise ratio is our bread and butter. Studying the minimum energy budget one can rely on to inform, to make inferences or to compute are not new problems, they are what we have been doing all along. But that is not where the novelty in these areas lies. If one explicitly considers the availability power to perform work as a stochastic process, pretty much all classical problems need to be revisited to revise what kind of performance can be guaranteed. Why is now power no longer given for granted? Because we do not want to use fossil fuel to
886
power-up information and cyber systems. That could help preserving more oxygen in the atmosphere which, by the way, it is not something to be cynical about. Just like availability of good fades or low congestion, inspired researchers to invent opportunistic forms of modulation and scheduling, understanding the laxity that exists in doing any work, starting from the work of transmitting or computing something, and ending with the work of our air conditioner or refrigerator, are exciting new areas to advance.
IEEE SIGNAL PROCESSING LETTERS, VOL. 19, NO. 12, DECEMBER 2012
Given that the roots of cybernetics are in the work of Norman Wiener, Rudolf Kalman and Claude Shannon, I think we are equipped to reinvent this field, if our legacy does not end up cutting our wings. ANNA SCAGLIONE, Fellow, IEEE, Editor-in-Chief Department of Electrical and Computer Engineering University of California at Davis Davis, CA 95616 USA
