Seminars & Webinars

Forthcoming Seminars/Webinars

2024

Link to the webinar: https://cassyni.com/events/NpjfAvCEN9vNrc71BZ6oMw

Abstract: “Hallucination” is a term used in the AI community to describe the plausible falsehoods produced by deep generative neural networks. It is often considered a negative, especially in relation with large language models or medical image reconstruction. Yet, in many computational photography applications, we rely on such hallucinations to create pleasing images. It often does not matter if all (or any) information was present in the real world if the produced falsehoods are visually plausible.
Starting from that premise, I will present our recent work on hallucinations in image reconstruction, image style creation, and texture synthesis, using different generative models such as GANs, diffusion networks, and neural cellular automata. With a nod to the dangers some of these hallucinations might pose, I will also briefly discuss our work on deep fake detection.

Bio: Sabine Süsstrunk is Full Professor and Director of the Image and Visual Representation Lab in the School of Computer and Communication Sciences (IC) at the Ecole Polytechnique Fédérale (EPFL), Lausanne, Switzerland. Her main research areas are in computational photography, applied machine learning, and computational image quality and aesthetics. Sabine is President of the Swiss Science Council SSC, Founding Member and Member of the Board (MOB) of the EPFL-WISH (Women in Science and Humanities) Foundation, MOB of the SRG SSR (Swiss Radio and Television Corporation) and the Swiss Study Foundation, and Co-Founder and MOB of Largo Films, Ltd. She received the IS&T/SPIE 2013 Electronic Imaging Scientist of the Year Award for her contributions to color imaging, computational photography, and image quality, and the 2018 IS&T Raymond C. Bowman and the 2020 EPFL AGEPoly IC Polysphere Awards for excellence in teaching. Sabine is a Fellow of ELLIS, IEEE, and IS&T.

Link to the webinar: https://us02web.zoom.us/j/86964016253?pwd=RbdA0RpFfb23mO0Dmqvtaxixd3Nnlz.1

Abstract:
The increasing aging population has led to a rise in pathological speech conditions such as dysarthria or apraxia of speech. These conditions are associated with neurological disorders such as Parkinson’s disease, Amyotrophic Lateral Sclerosis, or stroke. Diagnosing pathological speech in clinical practice involves time-consuming and expensive auditory-perceptual assessments by speech and language pathologists. To help alleviate this burden on the healthcare system, efforts within the research community are focused on developing automatic pathological speech processing approaches. This webinar will provide an overview of signal processing- and machine learning-based approaches for pathological speech processing. Examples of automatic detection and intelligibility assessment approaches will be discussed and insights on robustness to noise and security issues will be provided. 

Speaker:
Ina Kodrasi received the Master of Science degree in Communications, Systems, and Electronics from Jacobs University Bremen, Germany, in 2010, and the Ph.D. degree from the University of Oldenburg, Germany, in 2015. From 2010 to 2017, she was a research associate and a postdoctoral researcher at the Signal Processing Group of the University of Oldenburg, where she worked on multi-microphone dereverberation and noise reduction. Since December 2018, she has been with the Idiap Research Institute, Switzerland, where she leads the Signal Processing for Communication group. Her main research interests include signal processing, multi-channel processing, pathological speech processing, and machine learning. 

Seminars

Within the EURASIP Seminars program, financial support up to 1000 Euro from EURASIP can be obtained for inviting a speaker to deliver a lecture at the organiser’s institution. The invited speaker should have a strong link to EURASIP, such as for example EURASIP Fellows or Society Awardees.

Information for Prospective Seminar Organisers

If you are interested in organising a EURASIP Seminar, please contact the EURASIP Director for Technical Programs and Membership by email.

Webinars

Within the EURASIP Seminars program, financial support of 500 Euro can be obtained from EURASIP for the speaker to deliver high quality research webinars for the Signal Processing Community EURASIP. The invited speaker should have a strong link to EURASIP. A strong link can be evidenced by Membership of EURASIP, membership of a EURASIP Technical Area Committee, Awards of a EURASIP prize e.g. fellowship, publications in EURASIP Journals etc. To benefit from the financial support the invited speaker should be a member of EURASIP.

Information for Prospective Webinar Organisers

If you are interested in organising a EURASIP Webinar, please identify the most appropriate Technical Area Committee and email the Chair of the TAC outlining the topic for the webinar and the speaker.

Download here the consent form to share the recorded materials.

Past Seminars/Webinars

2024

Link to the recording: coming soon

Abstract:
In the areas of communications and compressed sensing, the demand for effective approximate Bayesian estimation techniques is paramount. Sparse channel modeling extends traditional model selection, enabling optimized models based on available training data. Compressed sensing techniques extend Linear Minimum Mean Squared Error (LMMSE) estimation by a hierarchical Bayesian formulation. In multi-user detection or blind channel estimation, going beyond LMMSE and Gaussian models represents a leap.
One of the approaches in the realm of approximate Bayesian estimation is Variational Bayes (VB), a relatively straightforward method. VB can be seen as an extension of the Expectation-Maximization (EM) technique to scenarios involving random parameters, thereby yielding not only point estimates but also approximate posterior distributions. Notably, while VB yields accurate means in Gaussian problems, it tends to underestimate variances significantly.
An even more refined technique for approximate Bayesian estimation is Expectation Propagation (EP). Both VB and EP share the underlying concept of minimizing the Kullback-Leibler Divergence (KLD), albeit with different sequencing of the true and approximating probability density functions. However, EP is a heuristic approach to minimizing a more desirable KLD, which is called the Bethe Free Energy (BFE). Exact alternating constrained minimization of the BFE leads to Belief Propagation (BP), whereas EP deviates from the alternating cost function and furthermore restricts approximating pdfs to be in an exponential family.
Taking a fresh look at alternating minimization of a KLD, the Central Limit Theorem leads to Gaussianity of the extrinsics in the marginal posteriors in moderate asymptotic settings. This in turn leads to what we call Gaussian Extrinsic Propagation, which sheds new light on characterizing performance beyond the loose Bayesian Cramer-Rao bound. Focusing on the Generalized Linear Model, assuming a n.i.i.d. (sign) statistical model for the measurement matrix allows asymptotically to find the variances without matrix inversions.
This leads to Approximate Message Passing (AMP) in which the Onsager correction term w.r.t. Jacobi iterations for solving the normal equations for the mean is related to the Componentwise Conditionally Unbiased MMSE estimation.
Reformulating AMP to correspond to alternating minimization of an asymptotic version of the BFE leads to a provably convergent algorithm. Alternating minimization becomes tricky in the presence of constraints and we shed some light on the desirable behavior of the Alternating Directions Method of Multipliers (ADMM) approach.

Speaker:
Dirk T.M. Slock received an EE degree from Ghent University, Belgium in 1982. In 1984 he was awarded a Fulbright scholarship for Stanford University, USA, where he received the MSEE, MS in Statistics, and PhD in EE in 1986, 1989 and 1989 resp. While at Stanford, he developed new fast recursive least-squares algorithms for adaptive filtering. In 1989-91, he was a member of the research staff at the Philips Research Laboratory Belgium. In 1991, he joined EURECOM where he is now professor. At EURECOM, he teaches statistical signal processing (SSP) and signal processing techniques for wireless communications. He invented semi-blind channel estimation, the chip equalizer-correlator receiver used by 3G HSDPA mobile terminals, spatial multiplexing cyclic delay diversity (MIMO-CDD) now part of LTE, and his work led to the Single Antenna Interference Cancellation (SAIC) integrated in the GSM standard in 2006. Recent keywords are multi-cell multi-user (Massive) MIMO, imperfect CSIT, distributed resource allocation, variational and empirical Bayesian learning techniques, large random matrix analysis, audio source separation, location estimation and exploitation. He graduated 40+ PhD students, leading to an edited book and 600+ papers. In 1992 he received one best journal paper award from IEEE-SP and one from EURASIP. He is the coauthor of two IEEE Globecom’98, one IEEE SIU’04, one IEEE SPAWC’05, one IEEE WPNC’16, one IEEE SPAWC’18, one SPAWC’23 and one CAMAD’23 best student paper award, and a honorary mention (finalist in best student paper contest) at IEEE SSP’05, IWAENC’06, IEEE Asilomar’06 and IEEE ICASSP’17. He cofounded the start-ups SigTone in 2000 (music signal processing products) and Nestwave in 2014 (Ultra Low-Power Indoor and Outdoor Mobile Positioning, acquired in 2022 by NextNav). He is a Fellow of IEEE and EURASIP. In 2018 he received the URSI France medal and he became a Knight of the French “Ordre des Palmes Académiques” in 2023.

Link to the recording: https://cassyni.com/events/LUiHRvuG5x9QtPh6DKsF3b

Abstract: Due to the increased popularity of augmented and virtual reality experiences, as well as 3D sensing for auto-driving, the interest in capturing high resolution real-world point clouds has grown significantly in recent years. Point cloud is a new class of signal that is non-uniform and sparse and this present unique challenges to the signal processing, compression and learning problems. In this talk, we present our multi-scale sparse convolutional learning and Graph Frourier Transform (GFT) based framework for large scale point cloud processing, with applications to the geometry and attributes super-resolution, and dynamic point cloud compression with latent space compensation. The architecture is memory efficient and can learn deep networks to handle large scale point cloud in real world applications. Initial results demonstrated that this framework achieved new state of the art results in geometry super-resolution, attributes deblocking and super-resolving, and dynamic point cloud sequence compression.

Bio: Zhu Li is a professor with the Dept of Computer Science & Electrical Engineering, University of Missouri, Kansas City(UMKC), and the director of NSF I/UCRC Center for Big Learning (CBL) at UMKC. He received his PhD in Electrical & Computer Engineering from Northwestern University in 2004. He was the AFRL summer faculty at the UAV Research Center, US Air Force Academy (USAFA), 2016-18, 2020-24. He was Senior Staff Researcher with the Samsung Research America’s Multimedia Standards Research Lab in Richardson, TX, 2012-2015, Senior Staff Researcher with FutureWei (Huawei) Technology’s Media Lab in Bridgewater, NJ, 2010~2012, Assistant Professor with the Dept of Computing, the Hong Kong Polytechnic University from 2008 to 2010, and a Principal Staff Research Engineer with the Multimedia Research Lab (MRL), Motorola Labs, from 2000 to 2008. His research interests include point cloud and light field compression, graph signal processing and deep learning in the next gen visual compression, remote sensing, image processing and understanding. He has 50+ issued or pending patents, 200+ publications in book chapters, journals, and conferences in these areas. He is an IEEE senior member, Associate Editor-in-Chief for IEEE Trans on Circuits & System for Video Tech, 2020~23, Associate Editor for IEEE Trans on Image Processing(2020~), IEEE Trans.on Multimedia (2015-18), IEEE Trans on Circuits & System for Video Technology(2016-19). He is the Chair of the IEEE Visual Signal Processing & Communication (VSPC) Tech Committee. He received the Best Paper Runner-up Award at the Perception Beyond Visual Spectrum (PBVS) grand challenge at CVPR 2023, Best Paper Award at IEEE Int’l Conf on Multimedia & Expo (ICME), Toronto, 2006, and IEEE Int’l Conf on Image Processing (ICIP), San Antonio, 2007.

Link to the recording: https://cassyni.com/events/TPm7xsSipzyNsysyWA6wmw

Abstract: The field of facial behavior analysis stands at the forefront of advancing intelligent systems, with applications that span clinical diagnostics, mental health assessment, customer service optimization, strategic business negotiations, and law enforcement. At the heart of this domain is the nuanced study of facial expressions, particularly micro-expressions, which are fleeting and subtle in nature, often lasting no more than half a second and characterized by minimal but significant facial movements. This research talk will delve into the intricacies of deciphering facial expressions, highlighting the complexities associated with micro-expressions due to their ephemeral nature and the low-amplitude facial cues they present. We will present our latest findings and methodologies that aim to enhance the accuracy and sensitivity of expression recognition technologies.

Bio: Wen-Huang Cheng is a University Distinguished Chair Professor in the Department of Computer Science and Information Engineering at National Taiwan University and a Visiting Professor at the Korea Advanced Institute of Science and Technology (KAIST). His current research interests include multimedia, computer vision, and machine learning. He has actively participated in international events and played significant leadership roles in prestigious journals, conferences, and professional organizations. These roles include serving as Editor-in-Chief for IEEE CTSoc News on Consumer Technology, Senior Editor for IEEE Consumer Electronics Magazine (CEM), Associate Editor for IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) and IEEE Transactions on Multimedia (TMM), General Chair for ACM MMAsia (2023), IEEE ICME (2022), and ACM ICMR (2021), Chair for IEEE CASS Multimedia Systems and Applications (MSA) technical committee, and governing board member for IAPR. He has received numerous research and service awards, including the Best Paper Award at the 2021 IEEE ICME and the Outstanding Associate Editor Award of IEEE TMM (2021 and 2020, twice). He is an IEEE Fellow, IET Fellow, and ACM Distinguished Member.

Link to the recording: https://youtu.be/xUGpT4_-C7A?si=SVHzK7eOs86vz69X

Abstract:
Code-based cryptography has been around for quite a while with the McEliece cryptosystem using the syndrome decoding problem in the Hamming metric considered one of the best understood cryptosystems with stable record of cryptanalytical advancement.  The best attacks are still message recovery attacks that use combinatorial methods against the underlying hard problem. 

In recent years, the quest for better performance has made the code-based scene much more colorful with abundance of new metrics, new hard problems and cryptographic constructions. As a result, the cryptanalytic methods are also more varied with algebraic methods becoming more relevant and more creative.

In this talk I will give an overview of algebraic attacks used in code-based cryptography. On a high level, such an attack involves modeling a hard problem or a cryptosystem as a  system of equations and then solving it. The challenge lies in finding the best possible algebraic model and the best possible solving method. I will go through several examples of algebraic modeling and solving of hard problems, decryption errors and cryptographic construction. I will further argue that often, a clever combination of algebraic and combinatorial methods yields the best results.

Speaker:
Dr. Simona Samardjiska is an assistant professor in post-quantum cryptography at the Digital Security group. Her expertise and research interests are in the mathematics of post-quantum cryptography (multivariate and code-based cryptography). She has been actively involved in the current NIST Post-Quantum standardization process as a principal submitter of the second-round candidate MQDSS and one of the submitters of MEDS in the new 4th NIST signature round. Besides being a designer, she has also contributed to the understanding of the security of several finalists and second-round candidates, by analyzing their classical security and resistance to Side-Channel attacks. She has published on several IACR conferences, journals and IEEE Symposiums and has been a program committee member of various cryptography-related conferences and workshops (PKC, CANS, SAC, PQcrypto, Africacrypt, Latincrypt). She is an editor of the EURASIP Journal on Information Security. She is an activist for gender balance and diversity in computer science.

Link to the recording: https://youtu.be/BMcNB1tW0YQ

Abstract:
The sixth generation (6G) of communication systems will be a key enabler of a plethora of new applications (e.g., autonomous driving, augmented reality, smart industry, etc.) that combine sensing, wireless communication, learning, and actuation, enabling the interaction among intelligent devices that collect and process considerable amounts of data. However, the stringent requirements imposed by the effectiveness of such applications need to face a bottleneck caused by resource scarcity, including spectrum, energy, computing, learning and inference capabilities. In this webinar, we present a fundamental paradigm shift for envisioned 6G systems given by the semantic and goal-oriented communications approach, which goes beyond the typical bit related metrics that are used today in system design and optimization, focusing instead on the recovery of the meaning conveyed by the transmitted bits and/or the effective fulfillment of the tasks motivating the exchange of information. We will present our recent results on the fundamental pillars of this system design, namely: (i) Goal-oriented data compression; (ii) Semantic representation learning and extraction via topological signal processing; (iii) Goal-oriented network resource optimization; (iv) Generative AI for semantic communications; and (v) RIS-enabled goal-oriented programmable wireless environments. The talk illustrates the practical advantages of this novel paradigm on several use-cases. Finally, we draw some conclusions and future research directions.

Speaker:
Paolo Di Lorenzo is an Associate Professor in the Department of Information Engineering, Electronics and Telecommunications, Sapienza University of Rome. He received the M.Sc. and Ph.D. degrees (magna cum laude) in Telecommunication Engineering in 2008 and 2012, respectively, both from Sapienza University of Rome, Italy. From September 2010 to April 2011, he held a research appointment in the Department of Electrical Engineering, University of California at Los Angeles. From May 2015 to February 2018, he was Assistant Professor in the Department of Engineering, University of Perugia, Italy. From March 2018 to February 2021, he was a Tenure Track Assistant Professor in the Department of Information Engineering, Electronics and Telecommunications, Sapienza University of Rome, Italy. His primary research interests lie in the areas of signal processing theory and methods, distributed learning and optimization, topological signal processing, and wireless communications. He is the technical manager of the SNS-JU European Project 6G-GOALS. He was the principal investigator of CNIT-Sapienza Research Unit in the H2020 European Project RISE 6G. He is recipient of the 2022 EURASIP Early Career Award, with citation “for contributions to the field of distributed signal processing, optimization, and learning over networks”. He has received three best student conference paper awards, which were sponsored by the IEEE signal processing society and the European Association for Signal Processing. He was also recipient of the 2012 GTTI award for the best Italian doctoral thesis in information and communication technologies. He serves as an Associate Editor for the IEEE Transactions on Signal Processing and, previously, for the IEEE Transactions on Signal and Information Processing over Networks and for the EURASIP Journal on Advances in Signal Processing. He is an IEEE senior member and a member of EURASIP.

Link to the recording: https://youtu.be/QIiHoVTRHvE?si=z9V1_JhIUp8fntYL

Abstract:
Decomposition of digital images into other basis or dictionaries than time or space domains is a very common and effective approach in image processing and analysis. Such a decomposition is commonly obtained using fixed transformations (e.g., Fourier or wavelet) or dictionaries learned from example databases or from the signal or image itself. In recent years, with the growth of computing power, data-driven strategies exploiting the redundancy within patches extracted from one or several images to increase sparsity have become more prominent. They have demonstrated very promising image restoration results. The question to pursue in this work is how to design such an adaptive transformation based on principles of quantum mechanics.
We investigate in detail a new approach of constructing such a signal or image-dependent bases inspired by quantum mechanics tools, i.e., by considering the signal or image as a potential in the discretized Schrodinger equation. To illustrate the potential of the proposed decomposition, denoising results are reported in the case of Gaussian, Poisson, and speckle noise and compared to the state-of-the-art algorithms. We further generalize our proposed adaptive basis by exploiting the data-driven strategy inspired by quantum many-body theory. Based on patch analysis, the similarity measures in a local image neighborhood are formalized through a term akin to interaction in quantum mechanics that can efficiently preserve the local structures of real images. The versatile nature of this adaptive basis extends the scope of its application to image-independent or image-dependent noise scenarios without any adjustment. Rigorous comparisons with contemporary methods demonstrate the superiority of the proposed algorithm regardless of the image characteristics, noise statistics, and intensity.
We show the ability of our approaches to deal with real-medical data such as clinical dental computed tomography image denoising and medical ultrasound image despeckling applications. Furthermore, we propose to enhance the spatial resolution of the 2D acoustic maps estimated from quantitative acoustic microscopy (QAM) imaging modality. QAM is an imaging system to form 2D quantitative maps of acoustic properties of soft tissues at a microscopic scale (<8 micrometers). Our custom-made QAM systems employ a 250-MHz or a 500-MHz single-element transducer to produce 2D maps with spatial resolutions of 7 or 4 micrometers, respectively. These state-of-the-art QAM systems remain inadequate for clinical investigations requiring even finer resolution, are costly to build, and require expert users. We propose a novel super-resolution algorithm by exploiting our off-the-shelf quantum-based adaptive denoiser (DeQuIP), leveraging the principles of quantum many-body theory. Adapting the recent breakthrough of regularization-by-denoising (RED) in image restoration, we impose this external DeQuIP denoiser as RED-prior combined with an analytical solution to handle the degradation operators for solving QAM super-resolution imaging. The efficiency of our proposed method is demonstrated by improving the resolution of experimental 2D acoustic impedance maps generated from the data acquired by 250-MHz and 500-MHz QAM systems. Results demonstrate that our RED algorithm permits the recovery of finer details and increases spatial resolution significantly. For example, a spatial resolution improvement of 40% was achieved when applied to acoustic impedance maps at 250-MHz, outperforming two other state-of-the-art methods, which only yielded 23-32% improvement.

Speaker:
Sayantan Dutta received the B.Sc. degree in mathematics from the University of Burdwan, Bardhaman, India, in 2016, the M.Sc. degree in mathematics from the Visva-Bharati University, Santiniketan, India, in 2018, and the M.S. degree in fundamental physics from the University of Tours, Tours, France, in 2019. He received Ph.D. degree in computer science from the University Paul Sabatier Toulouse 3, Toulouse, France, in 2023. He is currently a postdoctoral associate at the Department of Radiology, Weill Cornell Medicine, Cornell University, New York, USA.
His research interests include quantitative acoustic microscopy, quantum computing, quantum image processing, deep learning, and inverse problems, particularly denoising, deblurring, super-resolution, and compressed sensing.
In 2023, he was guest co-editor for the EURASIP JIVP special issue on “Recent Advances in Plug-and-Play Methods for Signal, Image and Video Processing: Theory, Algorithms, and Applications” and co-organizer of the special session on “Plug-and-Play Algorithms for Computational Imaging: Theory and Applications” at EUSIPCO.

Link to the recording

Abstract:
Most imaging inverse problems can be formulated as the reconstruction of an unknown image from noisy and potentially incomplete measurements. The traditional approach to reconstruction then consists of defining a cost function combining a data fidelity term and a regularisation term to compensate for the uncertainties associated with the measurements. Very often though, the optimization of such a cost function only provides point estimates, i.e., single restored images without useful associated uncertainty measures. This makes the use of the restored images within subsequent decision-making processes challenging. The Bayesian formalism allows for principled uncertainty management and quantification, but exact inference methods rarely scale well when the dimension of the problems increase. Thus, variational inference (VI) stands as a competitive alternative for uncertainty quantification at scale.
In this talk, we will discuss the application of Expectation-Propagation (EP) and variants, as an alternative to the more popular Variational Bayes (VB) methods used for approximate inference. In the context of image restoration, we will consider restoration from data corrupted by Gaussian and non-Gaussian noise and discuss how its modularity makes it a good candidate for a variety of inverse problems and priors/regularisation. We will review some of the current limitations of EP-based methods and will conclude with exciting avenues combining EP within large inference schemes.

Speaker:
Dr Altmann completed his Ph.D. within the Signal and Communication Group of the IRIT Laboratory in Toulouse, France, in 2013. In 2014, He was awarded a postdoctoral Fellowship by the Direction Générale de l’Armement (DGA, French Ministry of Defence) and joined Heriot-Watt University to develop computational methods for hyperspectral imaging and Lidar-based ranging applications. In 2017, he was also awarded a 5-year Research Fellowship by the Royal Academy of Engineering to develop new computational methods for low-illumination imaging and sensing.  Dr Altmann is the recipient of the EURASIP Early Career Award 2022 with Paolo Di Lorenzo. Since August 2022, he has been an Associate Professor at the School of Engineering and Physical Sciences, Heriot-Watt University. His research interests include Bayesian modelling and computation (Monte Carlo, variational inference) for imaging/sensing inverse problems, with applications in biomedical imaging, remote sensing and nuclear sciences.

Link to the recording

Abstract:
Brain MRI is becoming increasingly important in the study of both neurodevelopmental
and neurodegenerative processes due to its enormous capabilities in the study of brain anatomy
and function, with multiple applications in both research and clinical settings.
However, the inherently non-quantitative nature of most MRI sequences poses significant challenges
in the search for markers of pathology.
This webinar will provide an overview of the key factors involved in evaluating brain MRI
from a signal analysis perspective (e.g. signal registration, signal intensity normalisation,
signal denoising) and illustrate some examples from clinical and research fields.

Paolo Bosco
Dr. Paolo Bosco is a researcher at IRCCS Stella Maris Foundation, Pisa, Italy.
Since 2008 he focused his research interest in applying the methods of Physics in
medical image analyses and processing. He developed expertise in processing multi-modal imaging
of different organs and tissues and developed machine learning approaches for the detection of
novel imaging biomarkers of pathologies and conditions like breast cancer, Alzheimer’s disease and autism.

Link to the recording

Abstract:

Sound field reconstruction consists of estimating sound field properties from a set of spatio-temporal observations of an acoustic field. Such type of estimation is valuable in applications like spatial audio, active control of sound, VR/AR/XR, audio signal enhancement and transducer design – to name a few. There are however intrinsic challenges associated with measuring sound across space. Audible wavelengths span three orders of magnitude (17 m to 17 mm) and involve spatial sampling in three dimensions, resulting in a demanding problem. In this context, the talk focuses on the sensing, analysis, and reconstruction of sound fields over large spatial domains. We address the measurement of sound fields with microphone arrays, distributed sensors and remote acousto-optic sensing. We also discuss recent advances in signal processing and deep learning for sound field reconstruction, and demonstrate how models that capture generalizable physical properties of a sound field lead to better estimates than classical techniques that do not account for specific physical structure. As an outlook, we discuss how sound field analysis is approached in various domains of audio and acoustics and highlight some of the most promising avenues of current research in the field.


2023

Link to the recording

Abstract
Graphs are fundamental mathematical structures that are widely used in various fields for network data analysis to model complex relationships within and between data, signals, and processes. In particular, graph signals arise in many modern applications, leading to the emergence of the area of graph signal processing (GSP) in the last decade. GSP theory extends concepts and techniques from traditional digital signal processing (DSP) to data indexed by generic graphs, including the graph Fourier transform (GFT), graph filter design, and sampling and recovery of graph signals. However, most of the research effort in this field has been devoted to the purely deterministic setting. In this study, we consider the extension of statistical signal processing (SSP) theory by developing graph SSP (GSSP) methods and bounds. Special attention will be given to the development of GSP methods for monitoring the power systems, which has significant practical importance, in addition to its contribution to the enrichment of theoretical GSSP tools. In particular, we will discuss the following problems (as time permits): 1) Bayesian estimation of graph signals in non-linear models; 2) the identification of edge disconnections in networks based on graph filter representation; 3) the development of performance bounds, such as the well-known Cramér-Rao bound (CRB), on the performance in various estimation problems that are related to the graph structure; 4) the detection of false data injected (FDI) attacks on the power systems by GSP tools; 5) Laplacian learning with applications to admittance matrix estimation. The methods developed in these works use GSP concepts, such as graph spectrum, GSP, graph filters, and sampling over graphs.

Speaker
Tirza Routtenberg is an Associate Professor in the School of Electrical and Computer Engineering at Ben-Gurion University of the Negev, Israel. In addition, she was a William R. Kenan, Jr., Visiting Professor for Distinguished Teaching at the Electrical and Computer Engineering Department at Princeton University for 2022-2023. She was the recipient of four Best Student Paper Awards at international conferences. She is currently an Associate Editor of IEEE Transactions on Signal and Information Processing Over Networks and of IEEE Signal Processing Letters. In addition, she is part of the SPS Technical Directions Board Representative on the Education Board. Her research interests include statistical signal processing, graph signal processing, and optimization and signal processing for smart grids.

Link to the recording

Abstract:
The analysis and design of broadband multichannel systems typically involves convolutive mixing, characterised by matrices of transfer functions. Further, many broadband multichannel problems can be formulated using space-time covariance matrices that include an explicit lag variable and thus cross-correlation sequences as entries. This is in contrast to narrowband challenges, where the problem formulation relies on standard (i.e. constant) matrices; a rich set of solutions that are optimal in various senses can be reached from these formulations by matrix factorisations such as the eigenvalue or singular value decompositions. In order to extend the utility of such linear algebraic techniques to the broadband case, the diagonalisation or factorisation of matrices of functions is key.

In this webinar, I will show that such matrices are quite ubiquitous in multichannel signal processing, review some of the theory for their factorisations, and show how with such techniques broadband formulations and solutions directly generalise from their narrowband counterparts. I will sketch out a number of algorithms and illustrate their use in a few example applications such as beamforming, angle or arrival estimation, and signal compaction.

Speaker:
Stephan Weiss is a professor of signal processing at the University of Strathclyde. He received a Dipl.-Ing. degree from the University of Erlangen-Nuernberg in 1995, and a PhD in signal processing from the University of Strathclyde. Following academic appointments at Strathclyde (1998/99) and the University of Southampton (1999-2006), he has been back at Strathclyde since 2006. His interests include adaptive, multirate, and array signal processing;  In particular, with a number of collaborators, he has  championed polynomial matrix methods, the topic of this webinar, for the past two decades. He co-organised a number of events, such as EUSIPCO’09 and ICASSP’19 in Brighton, and served as head of Strathclyde’s Centre for Signal & Image Processing from 2015-2022.

Link to the recording

Abstract:
In recent years, information hiding methods have been increasingly applied in computer networks, especially by malware. This webinar starts with an introduction to network-level information hiding methods, including techniques of network covert channel research and selected methods of censorship circumvention. It highlights the recent advancements regarding a taxonomy for concealment and obfuscation methods (cf. [1,2] and https://patterns.ztt.hs-worms.de/) and provides an outlook on future steps to be undertaken in this context. Afterwards, the webinar highlights a novel concealment method that called a *history covert channel*. A history covert channel points to recently seen legitimate traffic to minimize the amount of transferred secret data [3] (as it only transfers a “pointer” to legitimate data). Further the webinar, presents a covert channel called epsilon-kappa-libur that circumvents highly-cited heuristics and two recent ML-based detection methods [4].

[1] Wendzel S, Caviglione L, Mazurczyk W, Mileva A, Dittmann J, Krätzer C, Lamshöft K, Vielhauer C, Hartmann L, Keller J, Neubert T, Zillien S (2022) A Generic Taxonomy for Steganography Methods, DOI: 10.36227/techrxiv.20215373

[2] Wendzel S, Caviglione L, Mazurczyk W (2023) Avoiding research tribal wars using taxonomies. IEEE Computer 56/1:93–96. IEEE, DOI: 10.1109/MC.2022.3218175

[3] Wendzel S, Schmidbauer T, Zillien S, Keller J (2022) Did You See That? A Covert Channel Exploiting Recent Legitimate Traffic, DOI: 10.48550/arXiv.2212.11850

[4] Zillien S, Wendzel S (2023) Weaknesses of popular and recent covert channel detection methods and a remedy. IEEE Transactions on Dependable and Secure Computing (TDSC)., DOI: 10.1109/TDSC.2023.3241451

Speaker
Steffen Wendzel is a professor of information security and computer networks at Hochschule Worms, where he is also the scientific director of the Center for Technology and Transfer. In addition, he is a lecturer at the University of Hagen. Before his professorship, he was a PostDoc at Fraunhofer FKIE in Bonn, where he led a research team on smart building security. He received his Ph.D. (Dr. rer. nat.) and his Habilitation (Dr. habil.) from the Faculty of Mathematics and Computer Science at the University of Hagen in 2013 and 2020, respectively. Website: https://www.wendzel.de

Link to the recording

Abstract:
The new era of healthcare digitization brings many advantages for patients and medical staff, but also transforms the medical institutions into a novel and valuable target of cybercriminals. Cybersecurity hazards involving healthcare institutions, such as big data breaches with stolen medical records, ransomware attacks, (Distributed) Denial of Service attacks followed by interruptions of various medical processes or insider threats, can be found nowadays in many headlines of world news agencies. The Picture Archiving and Communication Systems (PACS) infrastructure of modern hospitals deploys different standards and protocols for storing and transferring data between different modalities. The most used and important standard is DICOM (Digital Imaging and COmmunication in Medicine), which provide a framework for a diagnostically-accurate representation, processing, transfer, storage and display of medical imaging data. DICOM files combine medical media with patient/study/equipment/other metadata, while different transport mechanisms are defined for their exchange.
This webinar tries to demonstrate different security problems connected with the DICOM standard, such as the abuse of insecure or poorly-configured DICOM/PACS servers available on Internet; the possibility to access or perform different malicious manipulations on stored DICOM files, such as adding or removing evidence of medical conditions from volumetric medical scans by deep learning techniques, which can even deceive radiologists and state-of-the-art AI screening tools; abuse of information hiding techniques with the possibility to hide an executable code into a DICOM file, or to create covert channels useful for covert communication, privacy-leakage or data exfiltration, or to create and spread stegomalware; etc. In addition, security features currently used in DICOM are presented, together with crash course in DICOM.

Speaker:
Aleksandra Mileva is a Full professor at the Faculty of Computer Science, Goce Delcev University in Stip, N. Macedonia, where she is also the Head of the Laboratory of Computer Security and Computer Forensics. She received her PhD in Computer Science from the Ss. Cyril and Methodius University – Skopje in 2010. Her research interests include computer and network security, digital steganography, IoT protocols and information security, cryptography, computer forensics, and quasi-groups theory. She was with the management committee of two COST actions IC1201: BETTY and IC1306: Cryptography for Secure Digital Interaction, and she was Advisory Board member of H2020 SIMARGL project. She served as a guest editor for IEEE Internet of Things Journal, IEEE Security & Privacy, Journal of Universal Computer Science, co-chair of several conferences and workshops, and currently, she is a conference chair of the European Interdisciplinary Cybersecurity Conference (EICC) 2023. She is also a member of the editorial boards of Journal of Cyber Security and Mobility; Frontiers of Computer Science and Mathematics, Computer Science and Education. She is a co-author and developer of the NaSHA family of hash functions, which was the First Round Candidate of the NIST SHA-3 Competition (2007-2012). She was a coordinator of several scientific projects with PR of China, Portugal and Bulgaria. She has a certificate for Assessing and exploiting control systems and IIoT – Black Hat Edition. Mileva is a member of the EURASIP SAT on Biometrics, Data Forensics, and Security from 2019.

2018
  • Rémi Gribonval (France), Approximation with sparsely connected deep networks, 29 October 2018, Fornebu, Norway.
2017
  • Majid Ahmadi (University of Windsor, Canada), Mau-Chung Frank Chang (National Chiao Tung University, Hsinchu, Taiwan), Jiwu Huang (Shenzhen University, China) and more, RTSP 2017 EURASIP tutorial day, 10 – 11 July 2017, Bucharest, Romania.
  • Shlomo Shamai (Technion, Israel), Cloud and Fog Radio Access Networks: An Information Theoretic View, 15 March 2017, Berlin, Germany.
2016
  • Björn Ottersten (University of Luxembourg), Gerhard Fettweis (Technical University Dresden & Vodafone), Giuseppe Caire (Technical University Berlin), Petar Popovski (Aalborg University), Vincent Lau (Hong Kong University of Science and Technology), Robert Heath (University of Texas at Austin), David Gesbert (EURECOM), and Mérouane Debbah (Supélec & Huawei), 2016 Tyrrhenian International Workshop on Digital Communications (TIW16), 11 – 14 September 2016, Livorno, Tuscany, Italy.
  • Romain Couillet (Supelec) Merouane Debbah (Supelec & Huawei), Mario Figueiredo (IST) Georgios Giannakis (University Minnesota) Stephane Mallat (ENS) and Alain Rakotomamonjy (University Rouen), EURASIP Statistics, Optimization, and Signal Processing (STATOS) (link), 02 September 2016, Budapest, Hungary (right after EUSIPCO).
  • Mérouane Debbah (The role of Asymptotics in 5G), Luiz DaSilva (Spectrum and Infrastructure Sharing in Wireless Networks), Joint IEEE-EURASIP 3rd Spanish workshop on Signal Processing, Information Theory and Communication (SIC’16), EURASIP Liaison: Matilde Sánchez Fernández, 28-29 January 2016, Salón de Grados, Escuela Técnica Superior de Ingeniería (ETSI) Universidad de Sevilla, Spain.
2015
  • Dr. Henrique MALVAR, Prof. Julius O. SMITH, III and Kurt Werner, Prof. Ioan TĂBUȘ, Prof. Alessandro NERI, Prof. André KAUP, Prof. Erchin SERPEDIN, Prof. Wonyong SUNG, Prof. Yoshikazu MIYANAGA, Prof. David ALLSTOT, Prof. Guoan BI, RTSP 2015 – EURASIP Tutorial Day (following the program available at RTSP2015 Program), EURASIP Liaison: Corneliu Rusu, 06 July 2015, Grand Hotel Napoca, Cluj-Napoca, Rumania.
  • Prof. Gabriel Caffarena, Prof. Roberto Sierra, Prof. Ricardo Gómez, Prof. Pedro Guerra, Hardware Design of DSP Systems, EURASIP Liaison: Gabriel Caffarena, 29 April 2015, Universidad San Pablo-CEU, E.P.S., Urb. Monteprincipe, Boadilla del Monte, 28668, Madrid, Spain.
  • Evita Fotinea, Vassilis Katsouros, Third Greek Signal Processing Jam (SP-JAM3) in honour of George Carayannis, EURASIP Liaison: Constantine Kotropoulos, 20 April 2015, “Leonidas Zervas” amphitheatre, National Hellenic Research Foundation, Athens, Greece.
2014
  • R. Urbanke and F. Willems, Joint IEEE-EURASIP Spanish Seminar on Signal Processing, Communications and Information Theory, EURASIP Liaison: Matilde Sánchez (link), 10 – 11 December 2014, Univ. Carlos III, Madrid, Spain.
  • Prof. Mads Græsbøll Christensen, Statistical Parametric Speech Modeling EURASIP Liaison: Daniele Giacobello, 11 November 2014, Beats Electronics (Apple), 8600 Hayden Place, Culver City, CA 90232, United States.
  • Dipl.-Ing. Dr. techn. Wilfried Gappmair, Parameter estimation and synchronization in digital satellite receivers, 19 September 2014, CTTC Auditorium, Graz University of Technology, Austria.
  • Prof. Sergios Theodoridis, “Pattern Recognition: Principles and Beyond” (2nd IEEE SPS Italy Chapter Summer School on Signal Processing), EURASIP Liaison: Marco Carli (link), 11 July 2014, Frascati, Italy.
  • Anton Nijholt, “The 10th International Summer Workshop on Multimodal Interfaces” – SAM 2014 (Talk Information), 13 June 2014, Bilbao, Spain.
  • Satoshi Nakamura, “Multilingual automatic speech recognition” (4th International Workshop on Spoken Language Technologies for Under-resourced Languages – SLTU’14), EURASIP Liaison: Alexey Karpov (Report), 14-16 May 2014, SPIIRAS, St. Petersburg, Russia.
  • Prof. Miller Puckette (UCSD San Diego), Marc Leman (UGHENT), Christophe d’Alessandro (LIMSI CNRS Paris), Sébastien Roy (UMONTREAL), “CUTE 2014 masterclasses” (CUTE), EURASIP Liaison: T. Dutoit, 12 – 15 March 2014, Numediart Institute, 31 Bdv Dolez, Mons, Belgium.
  • Antonio Napolitano, “Doppler Effect on Almost-Cyclostationary Signals”, 12 – 14 February 2014, Technical University of Krakow, Poland.
2013
  • Abdelhak Zoubir, “Robust Statistics for Signal Processing”, Organizer: M. Barni (EURASIP Local Liaison) and A. Piva, 2 September 2013, Riotorto Livorno, Italy.
  • Anil K. Jain, “International Workshop on Bioinformatics and Forensics (IWBF 2013)”, Organizer: Paulo Correia (IST-IT, Lisbon, Portugal), (slides), 4 – 5 April 2013, Instituto Superior Técnico (IST) Congress Center, Lisbon, Portugal.
  • C. Pedreira, Madhav P. Desai, “IV EURASIP Seminar on Hardware Design of DSP Systems – Biomedical Signals: from neurons to custom processing hardware”, Organizer: Gabriel Caffarena (Univ. San Pablo), 6 March 2013, Escuela Politécnica Superior Universidad San Pablo CEU, Madrid, Spain.
  • Sara Rosenblum (University of Haifa, Israel), Eurasip Seminar on online handwriten analysis. Organizer: Marcos Faundez-Zanuy, 12 February 2013, Aula 103, Tecnocampus, Avda. Ernest Lluch 32, 08302 Mataró (Barcelona), Spain.
2012
  • Enrico Magli (Politecnico di Torino, Italy), Two-day workshop on Sparse Models and Machine Learning: “Compressed sensing for distributed communications”, Organizer: Rémi Gribonval, Local Liaison Officer: Aline Roumy, 15 October 2012, INRIA Rennes, Campus de Beaulieu, 35042 Rennes cedex, France.
  • Steve Young (University of Cambridge, UK), Recent Developments in Statistical Dialogue Systems (report, slides), 27 September 2012, Technische Universität Braunschweig, Institute for Communications Technology, Braunschweig, Germany.
  • Paul Van Dooren, Adaptive matrix factorizations and their use in tracking of dominant subspaces, Organizer: Laudadio Teresa, Local Liaison: Nicola Mastronardi (slides1, slides2), 27 – 31 Augustus 2012, Consiglio Nazionale delle Ricerche, via Amendola 122D, Bari, Italy.
  • A. Katsaggelos, A. Likas, I. Pitas, Y. Eldar, G. Giannakis, N. Kalouptsidis, C. Kotropoulos, 2nd Greek Signal Processing Jam, Organizer: Kostas Kotropoulos (slides, photos), Lecture videos are available by the host institution, 17 May 2012, Aristotle University, Thessaloniki, Greece.
  • Zabih Ghassemlooy (Northumbria University), Visible Light Communications (presentation), Organizer: Erich Leitgeb (poster of event), 11 May 2012, TU Graz, Austria.
  • Abdelhak M. Zoubir (Technische Universität Darmstadt), EURASIP funded seminar on “Robust Statistics for Signal Processing”, Organizer: Marc Moonen, 3 May 2012, Aula van de Tweede Hoofdwet, KU Leuven, Thermotechnisch Instituut, Kasteelpark Arenberg 41, Leuven – Heverlee, Belgium.
  • Karen Egiazarian and Vladimir Katkovnik (Tampere University of Technology), Block Matching 3-D (BM3D) image modeling and novel variational image reconstruction techniques, Organizer: COMLAB, (seminar report), 2 May 2012, Aula Magna Università degli Studi Roma TRE, Roma, Italy.
  • Sergios Theodoridis (U. of Athens), Adaptive Learning for Machine Learning and Signal Processing: An efficient Unifying Framework for Classification and Regression tasks (slides), Organizer: Markus Rupp, 11 April 2012, Technische Universität Wien, Vienna, Austria.
  • Prof. Mihaela van der Schaar, Repeated games and engineering economics (slides), Organizer: Beatrice Pesquet-Popescu, 15 – 16 March 2012, Telecom Paris-Tech, Paris, France.
  • Eduardo Boemo & Ruzica Jectiv, Hardware Design of DSP Systems: Low-Power Design (report), Organizer: Gabriel Caffarena (slides1, slides2), 8 February 2012, Universidad San Pablo – CEU, Madrid, Spain.
2011
  • Dr. Simon Dixon (U. of Queen Mary, London), Automatic Transcription: An Enabling Technology for Music Analysis (slides), First International Workshop on Folk Music Analysis: Symbolic and Signal Processing, Organizer: Aggelos Pikrakis, 19 May 2011, Old University of Athens, Plaka, Athens, Greece.
  • J.A. Lopez, G. Sutter, EURASIP Seminar on Hardware Design of DSP Systems, Organizer: Gabriel Caffarena, 27 April 2011, Universidad San Pablo CEU, Madrid, Spain.
  • Prof. Mihaela van der Schaar (UCLA), Shannon revisited:New separation principles for wireless multimedia (slides #1, slides #2), Organizer: Beatrice Pesquet-Popescu, 21 March 2011, Télécom ParisTech, Paris, France.
2010
  • Prof. Martin Vetterli (EPFL), Sampling Theory and Practice: 50 Ways to Sample Your Signal, Organizer: Akihiko K. Sugiyama (slides), images: image 1, image 2, image 3, 25 November 2010, Nara Women’s University, Nara, Japan (Map, link).
  • Dr. Anthony Vetro, (MERL), Secure Biometrics, Organizer: Paulo Lobato Correia (slides), 15 November 2010, IST-Portugal, Loures, Portugal.
  • Prof. Alan Willsky (MIT), Research in MIT’s Laboratory for Information and Decision Systems (LIDS) and in The Stochastic Systems Group, Organizer: Müjdat Cetin (slides, photos), 02 November 2010, Sabanci University, Tuzla/Istanbul, Turkey.
  • Dr. Josef Kittler (U. of Surrey), Information fusion in content-based retrieval from multimedia databases, Organizer: Ali Salah (Univ. of Amsterdam), slides, report, 04 August 2010, University of Amsterdam, Science Park, Amsterdam, the Netherlands.
  • Prof. Tim Davidson (U. of McMaster), Enriching the art of engineering design via convex optimization, Organizer: Mounir Ghogho (Leeds Univ.), presentation, 23 June 2010, Marrakech, Morocco.
  • Prof. Georgios B. Giannakis (U. of Minnesota), Distributed and sequential sensing of spatio-temporal spectra for cognitive radios, Organizer: Fulvio Gini (Univ. of Pisa), presentation, 14 June 2010, Elba Island, Italy.
  • Various Industry, Eurasip Seminar on Hardware Design of DSP Systems, Organizer: Gabriel Caffarena (Univ. San Pablo), 6 May 2010, San Pablo-CEU, Madrid, Spain.
2009
  • Riccardo Leonardi, The audio-visual message: structure and emotion: The talk was given within the context of GDR ISIS Groupement De Recherche ISIS “Scalable and cross-media indexing of multimedia content”, Organization: G. Quenot, J. Benois-Pineau, M. Cord, H. Bredin, GDR CNRS ISIS, announcement, presentation, 26 November, 2009, ENST, Paris, France.
  • Renato De Mori, Sergios Theodoridis, Nikos Sidiropoulos, Costantine Kotropoulos, Petros Maragos, Kostas Berberidis, Greek Signal Processing Jam, Presentations: Renato De Mori: Spoken Language Recognition and Understanding (presentation), Sergios Theodoridis: Adaptive Processing in a World of Projections (presentation), Nikos Sidiropoulos: Analyzing Data ‘Boxes’: Multi-way linear algebra and its applications in signal processing and communications (presentation), Costantine Kotropoulos: Music genre classification with multilinear and sparse techniques (presentation), Kostas Berberidis: Signal Processing & Communication Issues in Sensor Networks, image slide show, other material: Jam Poster, 17 October 2009, New Amphitheater, University of Athens, Athens, Greece.
  • Maria Sabrina Greco, EURASIP Seminar day on Radar Signal Processing: Hot topics and new trends, seminar material 1, seminar material 2, seminar material 3, photo 1, photo 2, 25 September 2009, Pisa, Italy.
  • Simon Haykin, Cubature Filters (pdf presentation), photo 1, photo 2, photo 3, 18 September 2009, Universidad Politecnica de Madrid, Madrid, Spain.