BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:5496@i2m.univ-amu.fr
DTSTART;TZID=Europe/Paris:20130321T140000
DTEND;TZID=Europe/Paris:20130321T150000
DTSTAMP:20241029T164016Z
URL:https://www.i2m.univ-amu.fr/evenements/n-chu-l2s-bayesian-approaches-v
 ia-sparsity-enforcing-priors-for-acoustic-source-imaging-with-robustness-s
 uper-resolution-and-wide-dynamic-range/
SUMMARY: (...): N. Chu (L2S): Bayesian approaches via sparsity enforcing pr
 iors for acoustic source imaging with robustness\, super resolution and wi
 de dynamic range
DESCRIPTION:: Bayesian approaches via sparsity enforcing priors for acousti
 c source imaging with robustness\, super resolution and wide dynamic range
 . By Ning Chu\, L2S.\n\nAcoustic imaging is a standard technique for mappi
 ng the positions and powers of acoustic sources with microphone arrays\, w
 hich provides the insight into the mechanisms and properties of acoustic s
 ources\, especially for acoustic comfort in automobile and aircraft indust
 ries. However\, acoustic imaging often involves the ill-posed inversions a
 nd becomes too time-consuming to obtain high spatial resolutions and wide 
 dynamic range.\n\nIn this talk\, we first improve the forward model by con
 sidering background noises\, as well as forward model uncertainty caused b
 y multipath propagation. Then we propose several robust super-resolution a
 pproaches for acoustic imaging via the following techniques:\n\n1 Sparsity
  constraint\,\n\n2 Sparsity regularization\,\n\n3 Bayesian inference via D
 ouble Exponential prior under joint Maximum A Posteriori (MAP) optimizatio
 n\,\n\n4 Variational Bayesian Approximation (VBA) via Student's-t priors f
 or both source powers and colored noises.\n\nFinally\, for efficient acous
 tic imaging\, we propose a shift-invariant 2D convolution model for acoust
 ic power propagation.\n\nOur approaches are compared with some of the stat
 e-of-the-art methods on simulations\, real data from wind tunnel experimen
 ts of Renault S2A\, as well as hybrid data. The main advantages of propose
 d approaches are robust to background (colored) noise\, wide dynamic range
  of acoustic powers\, super spatial resolutions\, efficient to use\, and n
 ot need for the prior knowledge of the source number or SNR. It is feasibl
 e to apply in near-field wideband monopole and extended source imaging bas
 ed on the 2D non-uniform microphone array for wind tunnel tests.\n\nDownlo
 ad slides
CATEGORIES:Séminaire,Signal et Apprentissage
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20121028T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR