EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2516. Space-Variant Total Variation boosted by learning techniques for subsampled imaging problems

Invited abstract in session MC-34: Optimization and learning for data science and imaging (Part III), stream Advances in large scale nonlinear optimization.

Monday, 12:30-14:00
Room: 43 (building: 303A)

Authors (first author is the speaker)

1. Davide Evangelista
University of Bologna

Abstract

In solving inverse problems with subsampled forward operators, a crucial aspect is choosing an efficient regularizer. The isotropic Total Variation (TV) functional is often used uniformly, to get gradient-sparse solutions, but it tends to remove small details and smooth edges. To address this, we propose a space-variant weighted isotropic TV regularization, whose weights can be determined by a neural network. This talk includes a rigorous theoretical analysis and presents experiments confirming the potential of the proposed approach to enhance the quality of undersampled CT reconstructions.

Keywords

Status: accepted


Back to the list of papers