EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
1795. A multiplicative components framework for joint correction and segmentation of magnetic resonance images
Invited abstract in session MA-34: Optimization and learning for data science and imaging (Part I), stream Advances in large scale nonlinear optimization.
Monday, 8:30-10:00Room: 43 (building: 303A)
Authors (first author is the speaker)
| 1. | Marco Viola
|
| School of Mathematical Sciences, Dublin City University | |
| 2. | Laura Antonelli
|
| Institute for high performance computing and networking (ICAR), Consiglio Nazionale delle Ricerche | |
| 3. | Valentina De Simone
|
| Mathematics and Physics, University of Campania "L. Vanvitelli" |
Abstract
The segmentation of Magnetic Resonance Images (MRIs) is a challenging task due to the artifacts introduced by the acquisition process, namely intensity inhomogeneity (also known as bias field) and noise (which follows a Rician distribution).
In this work we introduce a model called MICCT, based on the Multiplicative Intrinsic Components framework and Cartoon-Texture decomposition techniques, to perform simultaneous denoising and bias-field correction of an MRI.
The output of MICCT can be then be segmented with any state-of-the-art segmentation strategy.
We introduce an ADMM strategy to solve the nonlinear and nonconvex model associated with MICCT and analyse its theoretical convergence properties.
Finally, we present some numerical experiments showing the effectiveness and the competitiveness of the proposed approach (combined with a K-means strategy) in the segmentation of noisy MRI from the BrainWeb database.
Keywords
- Programming, Nonlinear
- Large Scale Optimization
- Auctions / Competitive Bidding
Status: accepted
Back to the list of papers