Skip to Main content Skip to Navigation
Documents associated with scientific events

Time-data tradeoff for the sparse and cosparse regularizations of physics-driven inverse problems

Srđan Kitić 1 Nancy Bertin 1 Rémi Gribonval 1
1 PANAMA - Parcimonie et Nouveaux Algorithmes pour le Signal et la Modélisation Audio
Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
Abstract : We investigate the computational performance of the sparse vs cosparse regularizations applied to physics-driven inverse problems, relative to the amount of measurements. Our results show that, despite nominal equivalence of the two models in the given context, the analysis-based optimization benefits from an increase in the volume of available data, while the synthesis one does not.
Document type :
Documents associated with scientific events
Complete list of metadata

Cited literature [15 references]  Display  Hide  Download

https://hal.inria.fr/hal-01147438
Contributor : Srdan Kitic <>
Submitted on : Thursday, June 25, 2015 - 1:41:22 PM
Last modification on : Thursday, January 7, 2021 - 4:19:35 PM
Long-term archiving on: : Tuesday, April 25, 2017 - 7:09:02 PM

File

SPARS15.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01147438, version 2

Citation

Srđan Kitić, Nancy Bertin, Rémi Gribonval. Time-data tradeoff for the sparse and cosparse regularizations of physics-driven inverse problems. SPARS15 - Signal Processing with Adaptive Sparse Structured Representations, Jul 2015, Cambridge, United Kingdom. ⟨hal-01147438v2⟩

Share

Metrics

Record views

1001

Files downloads

194