Modelling Data with both Sparsity and a Gaussian Random Field: Application to Dark Matter Mass Mapping in Cosmology
- Resource Type
- Conference
- Authors
- Themelis, Konstantinos E.; Lanusse, Francois; Jeffrey, Niall; Peel, Austin; Starck, Jean-Luc; Abdalla, Filipe B.
- Source
- 2018 26th European Signal Processing Conference (EUSIPCO) Signal Processing Conference (EUSIPCO), 2018 26th European. :376-379 Sep, 2018
- Subject
- Bioengineering
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Signal Processing and Analysis
Convergence
Signal processing algorithms
Covariance matrices
Shape
Optimization
Noise measurement
Data models
- Language
- ISSN
- 2076-1465
In this paper we present a novel method for dark matter mass mapping reconstruction from weak gravitational lensing measurements. The crux of the proposed method lies in a new modelling of the matter density field in the Universe as a mixture of two components: a) a sparsity-based component that captures the non-Gaussian structure of the field, such as peaks or halos at different spatial scales; and b) a Gaussian random field, which is known to well represent the linear component of the field. This new model represents the distribution of matter in the universe much better than previously proposed models. We have developed a new algorithm that also takes into account the experimental problems we meet in practice, such as a non-diagonal covariance matrix of the noise or missing data. Experimental results on simulated data show that the proposed method exhibits improved estimation accuracy compared to state-of-the-art methods.