lu.se

Matematikcentrum

Lunds universitet

Denna sida på svenska This page in English

Maria Sandsten

Professor in Mathematical Statistics-Statistical Signal Processing

My research interests focus on time-frequency and spectrum analysis of especially non-stationary stochastic processes. The main applications are biomedical signals and also signals and sound/speech connected to humans and animals. 

Examples of research projects

Time-Frequency Localization

The reassignment technique is used to increase localization for signal components in the time-frequency representation. The technique gives perfect localization for infinite linear chirp-signals, impulses and constant frequency signals but not for short non-stationary signals.  Based on the spectrogram using a Gaussian window we propose a scaled reassignment that gives perfect localization for a Gaussian function. To the right an example of two Gaussian components clearly visualized by the scaled reassignment (ScRe-Spect). The method has been applied to estimation of dolphin echo-location signals in a collaboration with Josefin Starkhammar, division of Biomedical Engineering, Lund University.

Modeling and classification of bird song

The project is a part of the e-science collaboration,  eSSENCE, and is performed together with Dennis Hasselquist, Bengt  Hansson and Maja Tarka at the Department of Biology, Lund University. The project applies to general bird song but is specifically focused on the song of the Great Reed Warbler.

Estimation of chirp-HRV power

The project focus on robust estimation of Heart Rate Variability power and in the collaboration with Peter Jönsson, Department of Laboratory Medicine, Lund University, we have established a special measurement method where the patient should breath according to a metronom with increasing frequency. Using robust time-frequency methods the power is estimated from the resulting non-stationary HRV-signal. 

CV

Master of Science in Electrical Engineering, Lund University,1989, and Ph.D. in Signal Processing, Lund University, 1996. Docent in Signal Processing, 2004 and Professor in Mathematical Statistics with speciality in Statistical Signal Processing, 2010.

Main supervisor of 4 PhD students, assistant supervisor of 3 PhD students and supervisor of 1 Post Doc.

Member of the board of recruitments, LTH, 2013-2015. Vice coordinator for the program Engineering in Mathematics, LTH, 2013-2015. Head of division Mathematical Statistics, 2008-2012. Member of the board of the strategic research area E-science at LTH, 2013-2014. Member of the board of Lunarc, 2012-2014.  

Larger grants: Strategic research for E-science-Essence, 'Classification tool for bird singing', 1440 kkr, 2012-2014. VR-621-2011-5813, 2400 kkr, 2012-2014, VR-621-2007-6264, 750 kkr, 2008. VR-621 2004-4692, 1860 kkr, 2005-2007. Researcher in technical science, 'särskild forskare', VR-622-2003 602, 5700 kkr, 2004-2009. VR-621-2001-3120, 1200 kkr, 2002-2003, (co-applicant). VR-621-2001 3119, 1000 kkr, 2002-2003. TFR-271-2000-605, 490 kkr, 2001. TFR-271-1997-627, 1400 kkr, 1998 2000.

Publications of more than 20 international journal papers and over 30 conference papers. Co-author of 6 patents. 

Associate editor for IEEE Trans. on Signal Processing, 2008-2010. Regular referee for IEEE Trans. on Signal Processing, Elsevier Signal Processing, IEEE Trans. on Biomedical Engineering, IEEE Trans. on Speech and Audio Processing, Elsevier Medical Engineering and Physics. 

Publications

 

Email: sandsten@maths.lth.se

Room: 127

Phone (Office):
+46 46 22 249 53

Phone (Dept.):
+46 46 22 285 50

Fax:
+46 46 22 246 23

Internal postal system: 6

Visitors address:
Sölvegatan 18

Postal address: 
Mathematical Statistics
Centre for Mathematical Sciences
Lund University
Box 118
SE-22100 Lund
Sweden

Current teaching

Stationary stochastic processes 

Stationary and Non-Stationary Spectral Analysis

Suggestions of master thesis project:

Classification of bird song syllables using low-rank approximations of time-frequency images

Feature extraction for dialect classification

Cross-spectral analysis of HRV data connected to work related stress

For more information on the projects contact Maria Sandsten

Arbetsgång i examensarbetet