Events

13. June 2024, 15:00 until 17:00
AKOR Seminar: A convex Formulation for Total-Variation Parameter Learning, Enis Chenchene, University of Vienna
Conference
We present a new approach for data-driven tuning of regularization parameters for total-variation denoising.
The proposed approach can be understood in connection to a weak optimal transport problem between the distributions of clean and noisy data and admits a manageable convex formulation that can be solved efficiently with a new conditional-gradient-type method. We show numerical experiments and open avenues for promising extensions.