Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Parameter estimation of Gaussian functions using the scaled reassigned spectrogram

Brynolfsson, Johan LU and Sandsten, Maria LU (2015) 23rd European Signal Processing Conference, 2015 p.988-992
Abstract
In this paper we suggest an improved algorithm for estimation of parameters detailing Gaussian functions and expand it to handle linear combinations of Gaussian functions. Components in the signal are first detected in the spectrogram, which is calculated using a Gaussian window function. Scaled reassignment is then performed using a set of candidate scaling factors and the local Renyi entropy is used to measure the concentration of each component using every candidate scaling factor. Exploiting the fact that a Gaussian function may be perfectly reassigned into one single point given the correct scaling, one may identify the parameters detailing the Gaussian function. We evaluate the algorithm on both simulated and real data.
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
Signal Processing Conference (EUSIPCO), 2015 23rd European
pages
5 pages
publisher
IEEE - Institute of Electrical and Electronics Engineers Inc.
conference name
23rd European Signal Processing Conference, 2015
conference location
Nice, France
conference dates
2015-08-31 - 2015-09-04
external identifiers
  • scopus:84963969130
DOI
10.1109/EUSIPCO.2015.7362531
language
English
LU publication?
yes
id
65c0bf17-2d60-4d86-af3f-d56eedfe9db5 (old id 8310641)
date added to LUP
2016-04-04 10:47:24
date last changed
2022-01-29 20:47:49
@inproceedings{65c0bf17-2d60-4d86-af3f-d56eedfe9db5,
  abstract     = {{In this paper we suggest an improved algorithm for estimation of parameters detailing Gaussian functions and expand it to handle linear combinations of Gaussian functions. Components in the signal are first detected in the spectrogram, which is calculated using a Gaussian window function. Scaled reassignment is then performed using a set of candidate scaling factors and the local Renyi entropy is used to measure the concentration of each component using every candidate scaling factor. Exploiting the fact that a Gaussian function may be perfectly reassigned into one single point given the correct scaling, one may identify the parameters detailing the Gaussian function. We evaluate the algorithm on both simulated and real data.}},
  author       = {{Brynolfsson, Johan and Sandsten, Maria}},
  booktitle    = {{Signal Processing Conference (EUSIPCO), 2015 23rd European}},
  language     = {{eng}},
  pages        = {{988--992}},
  publisher    = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
  title        = {{Parameter estimation of Gaussian functions using the scaled reassigned spectrogram}},
  url          = {{http://dx.doi.org/10.1109/EUSIPCO.2015.7362531}},
  doi          = {{10.1109/EUSIPCO.2015.7362531}},
  year         = {{2015}},
}