You are here: Home > About Us > People > Sabine Van Huffel

Sabine Van Huffel

Sabine Van Huffel

Research

Research – fundamental/theoretical as well as application oriented- is performed in the domain of (multi)linear algebra, (non)linear signal analysis, machine learning and Artificial Intelligence (AI) with special focus to the development of numerically reliable and robust algorithms for empowering medical decision support and smart patient monitoring. In this domain the group has built up an international reputation.
Highlighted topics are:

Topics:

  • Total Least Squares (TLS) fitting  combines effectively statistical and numerical methodologies for dealing with errors in linear parameter estimation problems; Her contributions involve numerical issues, computational efficiency, statistical value, and valuable, creative applications mainly in biomedicine. Her book (The Total Least Squares Problem, SIAM, 1991, 300 pages, about 2055 scholar-google citations) with J. Vandewalle became the standard book for TLS and received excellent reviews. She organized 4 workshops on TLS and errors-in-variables modeling, attracting 70 scientists from diverse disciplines to Leuven, Belgium (1991, 1996, 2001, 2006), which resulted in 2 edited books (SIAM,1997) and (Kluwer, 2002)1 and 2 special issues (Signal Processing 2007, Computational Statistics & Data Analysis 2007).
  • Numerical tensor algorithms for biomedical multimodal and multichannel data processing including applications in EEG-fMRI and MRSI-MRI integration. A major leap forward here was achieved in 2014-2019 when Sabine Van Huffel was awarded an ERC advanced grant BIOTENSORS (no. 339804) of 2,5 milion euro, entitled "Biomedical Data Fusion using Tensor based Blind Source Separation". Together with her team, she developed a general functional framework for solving tensor based blind source separation (BSS) problems in biomedical data fusion, using tensor decompositions (TDs) as basic core. These TDs allow the extraction of fairly complicated sources of biomedical activity from fairly complicated sets of uni- and multimodal data. Significant progress has been made on the development of algebraically and numerically well-founded tensor techniques for BSS, including the Canonical Polyadic Decomposition (CPD) and Block Term Decomposition (BTD) and their extensions involving block terms, various constraints, source models and heterogeneous coupled data sets. Tensorlab efficiently implements the above algorithms (see http://www.tensorlab.net/). Its user friendliness has improved significantly by simplifying model construction and adding demos and Matlab-based GUIs for CPD and data compression use and tensor-based biomedical BSS. These improvements greatly facilitate non-expert usage and allow to face current grand challenges in biomedical data fusion. Various tensor-based approaches, all efficiently implemented in Tensorlab, have been applied successfully in diverse biomedical BSS problems in ECG, EEG, and MRSI signal processing. The added value of BTD versus CPD to detect seizure patterns evolving in space or frequency was proven. In addition, the power of tensor-based approaches in abnormal heartbeat detection using ECG has been shown. The added value of coupled matrix-tensor factorisations (CMTF) is shown in epileptic EEG-fMRI: the sources corresponding to the onset (spike) and propagation (slow wave) can be separated, enabling to better localize the seizure onset zone. By adding a Toeplitz factor, CMTFs allow to perform both BSS (estimation of spatial-temporal-spectral activations) but also blind system identification (estimation of neurovascular coupling or transfer function between modalities). More information on the research, publications and software is available on the project website
  • In decision making, a variety of powerful algorithms have been developed based on least squares (LS) support vector machines and neural networks (NN), ranging from classical multilayer perceptrons to convolutional neural networks (CNNs) as used in deep learning. Significant contributions to medical diagnostics, e.g. seizure detection, neonatal monitoring, ovarian & brain cancer diagnosis, have been shown. Recently, our epileptic seizure detection pipeline fusing multiview deep NNs, was awarded the first place in the Neureka 2020 Epilepsy Challenge, see https://neureka-challenge.com/results/, and best paper award.
  • Medical decision support now requires an ingenious combination of all above-mentioned tools, starting from an adequate pretreatment of the data (e.g. artefact removal), feature selection, pattern recognition, decision making, up to their embedding into user-friendly user interfaces. Many of these tools have been implemented in Matlab and are freely available.
  • Applications are oriented towards both ambulatory and in-hospital smart patient monitoring and decision support based on (multimodal) biomedical data fusion (e.g. multichannel EEG, ECG, EMG, accelerometry, polysomnography, audio, video ...). In close collaboration with UZ Leuven, multiple decision support tools have been developed in various fields such as epilepsy monitoring, sleep and neonatal brain monitoring, cancer diagnostics, Magnetic Resonance Spectroscopic (MRS) and MRS Imaging (MRSI), cardiovascular dynamics and heart-rate variability analysis, event-related potential (ERP) analysis, and stress monitoring. In particular, the widely used AMARES[1] software is nowadays a standard technique for long echo-time MRS quantification.
    [1] L. Vanhamme, A. Van Den Boogaart,  S. Van Huffel, Improved  method for accurate and  efficient quantification of MRS data with use of  prior knowledge. J. Magn.  Resonance, vol.129, no.1, p.35-43,  1997,  has  1649 scholar-google citations

University Degrees:

  • July 1981: Master of Science in Computer Science Engineering, KU Leuven, Univ. of Leuven, Belgium.
  • September 1981: Licensed teacher in Higher Secondary Education, KU Leuven, Univ. of Leuven, BE.
  • July 1985: Postgraduate degree in Biomedical Engineering, KU Leuven, University of Leuven, Belgium.
  • June 1987: Ph.D. in Electrical Engineering, Dept. of Electrical Engineering, KU Leuven, Summa cum laude.

Professional Experience:

  • Oct. 1 2020: Professor Emerita with duties: supervision of Ph.D. research, coordination of ongoing projects and follow-up of the implementation of the new Biomedical Engineering Educational programmes
  • 2014-present: Distinguished professor at Eindhoven University of Technology, Eindhoven, The Netherlands
  • 2002-2020: Full professor, Department of Electrical Engineering (ESAT), Division STADIUS, KU Leuven, and head of biomedical data processing research group (including approximately 3 staff members, >30 PhD students, 4 postdocs).
  • 1998-2002: Professor (part-time till Oct.1,2000), Dept. of Electrical Engineering (ESAT), KU Leuven, BE
  • 1999–2000: Research director Research Foundation Flanders (F.W.O.), BE
  • 1995-1998: part-time associate professor, Department of Electrical Engineering (ESAT), KU Leuven, BE
  • 1991-1999: research associate Research Foundation Flanders (F.W.O.), BE
  • 1993-1995: part-time assistant professor, Department of Electrical Engineering (ESAT), KU Leuven, BE
  • 1987-1991: postdoctoral researcher Research Foundation Belgium (F.W.O.), BE
  • 1983-1987: PhD student, Department of Electrical Engineering (ESAT), KU Leuven, Belgium
  • 1981-1983: scientific researcher, Faculty of Medicine, KU Leuven, BE

Visiting Positions:

  • Oct.-Nov. 2000: Guest professor, Dept. Computer Sciences, Stanford University, CA (host: Gene Golub)
  • May-June 2004: Guest professor, Dept. Systems & Control, Uppsala University, Sweden (host: P. Stoica)
  • June-Aug. 1992: Visiting fellow, ARHPRC, Univ. of Minnesota, Minneapolis, U.S.A. (host: Haesun Park)
  • April-May 1993: Visiting scientist, Computer Science Dept., University of Minnesota, Minneapolis, U.S.A.

Positions in International Societies:

  • 2000-2020: Head BIOMED (Biomedical data processing) research group, ESAT: 3 staff members, 4 postdocs, > 30 PhD students.
  • Programme Director of Biomedical Engineering Education at KU Leuven, KU Leuven from August 1, 2016, till August 2020.
  • Honorary Doctorate at Eindhoven University of Technology: April 25, 2013.
  • Distinguished Professor at Eindhoven University of Technology: January 1, 2014 – January 1, 2022
  • 2005-2009: Rectorial advisor for equal opportunities and Diversity. In this capacity she has taken various initiatives. for attracting more female students to engineering.
  • Member of the SIAM Fellows Selection Committee (2017-2018)
  • Member of the IEEE Fellows Selection Committee (2016-2020)
  • Chair (2011-2013) & Panel member (since 06) Commission Informatics & Knowledge Technology FWO
  • Member Householder Prize Committee ('02-'11): selects 3-yearly best PhD in numerical linear algebra
  • Member Barco and IBM Prize committees (annual selection best master thesis and best PhD thesis)
  • Cofounder and treasurer IEEE-EMBS Benelux Chapter (2005-2020)
  • President (2011) and steering committee member (since 2006) of ISOTT
  • Member National Committee on Biomedical Engineering, IEEE, ISMRM, ESMRMB, EURASIP, SIAM
  • Chair Numerics-in-Control NICONET International Society & editor NICONET reports (1998-2005).
    The NICONET Society was the result of international cooperation among leading numerical algebra and control numerics experts within an EU thematic network, also called NICONET, which continued afterwards for many years till today. S. Van Huffel has been the chairperson of NICONET from 1991 till 2005. Its main activity was the development of quality control-application software that performs efficiently and reliably on modern computers using the subroutine library SLICOT (Subroutine Library in Systems and Control Theory) as example. The SLICOT library was licensed to Mathworks. Its routines (in total more than 570 at present) are partly incorporated in Matlab from release 7 on and improve the computational efficiency of equivalent Matlab functions significantly, often by orders of magnitude, especially in toolboxes for system identification and control. See http://slicot.org/
  • Member Editorial board of EURASIP J. Signal Process. and Bioinformatics (since 2005), Springer J. Signal, Image and Video Process. (since 2006), Numer. Lin. Alg. with Appl. (2004-2010), EURASIP J. on Signal Process. (2003 -2005), Numer. Algor. (1995-2012), SIAM J. Matrix Anal. and Appl. (1996-2005).

Invited presentations:

55 in total in 2004-2021. The most important ones since 2015 are:

  • "Machine learning in neonatal brain monitoring", Invited webinar, 12th International Newborn Brain conference Series, Digital Diagnostics for the Developing Brain, May 20, 2021. More than 200 participants expected.
  • Challenges in neonatal monitoring. Invited Seminar, Stadius Center for Dynamical Systems, Signal Processing, and Data Analytics, Department of Electrical Engineering (ESAT), KU Leuven, Belgium, December 2, 2020. More than 80 participants.
  • "Challenges in neonatal EEG monitoring" and "Cerebral autoregulation techniques using NIRS". Two Invited talks, 9th International Summer School in Biomedical Engineering: Noninvasive dynamic brain imaging in infants, children, and adults, August 05 - 11, 2020, Erfurt, Germany. More than 60 participants.
  • Blind source separation. Invited speaker, ANT Neuro meeting, 29th edition, January 15-18, 2020, Beaunes, France. More than 150 participants.
  • Challenges in Neonatal Brain Monitoring, Keynote Talk, 41st Annual Symposium of the IEEE Engineering in Medicine and Biology Society, Berlin, July 23-27, 2019
  • Biomedical data fusion using tensor-based blind source separation. Keynote Talk, EURASIP Summer School on Tensor-based Signal Processing, Leuven, August 27-31, 2018.
  • Classification Tools for MRS of Cancer. Invited speaker, Joint meeting ISMRM-ESMRMB, Paris, France, June 16-21, 2018. 
  • The Power of Low Rank Tensor Approximations in Smart Patient Monitoring. Keynote Speaker, European Signal Processing Conference (EUSIPCO 2017), Kos, Greece, August 28-September 2, 2017.
  • Automated MR Spectra Classification for diagnostic support: tools and applications. Invited speaker, Annual meeting ESMRMB, Vienna, September 29-October 1, 2016.
  • The Power of Low Rank Matrix and Tensor Approximations in Smart Diagnostics. Workshop on Low Rank Optimisation and Applications, June 8-12, 2015, Haussdorf Center for Mathematics, Bonn, DE.
  • The Power of Matrix and Tensor decompositions in ECG Monitoring. International Conference on Basic and Clinical Multimodal Imaging (BaCi), September 1-4, 2015, Domkerk, Utrecht, NL.
  • The Power of Matrix and Tensor decompositions in Smart Patient Monitoring. Keynote talk SIAM Conference on Computational Science and Engineering (CSE15), March 14--18, 2015, Salt Lake City, Utah (1800 participants)

Organisation of international conferences:

16 in period 2003-2020. The most important ones are listed below.

  • Chair IEEE-EMBS Benelux chapter symposium ``Artificial Intelligence In Healthcare’’, Leuven, Belgium, November 28-29, 2019.
  • Chair EURASIP Summer school on Tensor-based Signal Processing, August 27-31, 2018
  • Chair International Workshop Tensor Decompositions and Applications, Jan. 18-22, 2016
  • Member organizing Committee SIAM Conf. Applied Linear Algebra (LA15), Atlanta, USA, Oct.26-30, 2015.
  • Member organizing committee Householder XIX Symposium, Spa, Belgium, June 8-13, 2014.
  • Chair (Bruges, BE, August 19-24, 2012), member programme committee (Ascona, July 18-23, 2010) & steering committee (since 2006) of annual meeting Int. Society on Oxygen Transport to Tissue (ISOTT).
  • Programme chair Int. Conf. on Bio-inspired Systems and Signal Processing, subconf. of 5th Int. Joint Conf. on Biomedical Eng. Systems and Technologies (BIOSTEC), Algarve, Portugal, Feb.1-4, 2012.
  • Member programme and Organizing Committee European Congress of the International Federation for Medical and Biological Engineering (MBEC 2008), Antwerp, Belgium, Nov. 23-27, 2008
  • Chair 4th Int. workshop "TLS and Errors-in-Variables Modeling", Leuven, BE, Aug. 21-23, 2006.
  • Co-chair 4th workshop of ERCIM Working Group on "Matrix Computations and Statistics" jointly with 1st Int. workshop on "Numerical Linear Algebra and its Applications", Monopoli, Sept.22-24,2003.
  • Co-organizer XXI summer School of Computational Mathematics, Monopoli, Italy, Sept. 15-20, 2003.

Current research team

Pooya Ashtari

Biography

  • International Study Programme in Statistics, KU Leuven: propaedeutic course in Calculus, linear algebra and statistics & advanced course in Errors-in-Variables Modeling (1998-2007).
  • M. S. Electrical Engineering, KU Leuven: 2 courses in Biomedical Data Processing (2002-2007).
  • M.S. Biomedical Engineering, KU Leuven: “Biomedical Data Processing” (2002-2020), "Biomedical Data Processing, Part II" (2010-2020), “Design in Medical Technology” (2007-2020), "Industrial internship" (2010-2020).
  • M.S. Bioengineering, KU Leuven: "Human Health Data Processing" (2012-2020).
  • PhD courses: "numerical methods in biomathematics" at Stanford University, USA (fall 2000), Graduate School of Systems and Control, KU Leuven (winter 2000), Uppsala University, Sweden (May-June 2002) and at the ESAT Laboratory, KU Leuven (January 2005).

(Co)supervisor of 71 finished PhDs and 11 ongoing PhDs: mostly all interdisciplinary in cosupervision with medical colleagues. Postdocs give support through daily supervision. In addition, she mentored more than 30 postdocs since 1994. Graduates include research lab/faculty members in respectable institutions (TU Delft, Univ. of Oxford, Univ. Georgia Tech, Univ. Southampton, Univ. Kiev, Univ. of Sydney, Univ. Picardie, Univ. Wales, Univ. Jakarta, Univ. of Oldenburg, IMEC, CNR Bari, Sony, Philips Medical Systems, Alcatel). Ivan Markovsky, former PhD and postdoc and now professor at Brussels free University (VUB) in Belgium, got 2nd Householder prize (honourable mention) in 2008.

Finished projects

    Awards

    Editorship

    Membership

    Publications

    301 Moved Permanently

    301 Moved Permanently

    Contact information