Concept
Anisotropic diffusion
Publications associées (57)
Maximum likelihood bolometry for ASDEX upgrade experiments
Bolometry is an essential diagnostic for calculating the power balances and for the understanding of different physical aspects of tokamak experiments. The reconstruction method based on the Maximum Likelihood (ML) principle, developed initially for JET, h ...
Bristol2023The effects of a spur dike location in a 90° sharp channel bend on flow field: Focus on anisotropy degree and anisotropy nature
In this research, the flow features around a spur dike located in a 90˚ sharp channel bend have been studied experimentally in detail. Results showed that the effects of the spur dike on upstream sections increased by increasing α (spur dike location from ...
2023Annihilation Filter Approach For Estimating Graph Dynamics From Diffusion Processes
Pascal Frossard, Arun Venkitaraman
We propose an approach for estimating graph diffusion processes using annihilation filters from a finite set of observations of the diffusion process made at regular intervals. Our approach is based on the key observation that a graph diffusion process can ...
IEEE2022Image Denoising with Control over Deep Network Hallucination
Sabine Süsstrunk, Majed El Helou, Qiyuan Liang
Deep image denoisers achieve state-of-the-art results but with a hidden cost. As witnessed in recent literature, these deep networks are capable of overfitting their training distributions, causing inaccurate hallucinations to be added to the output and ge ...
Society for Imaging Science and Technology (IS&T)2022Coarse graining and large-N behavior of the d-dimensional N-clock model
We study the asymptotic behavior of the N-clock model, a nearest neighbors ferromagnetic spin model on the d-dimensional cubic epsilon-lattice in which the spin field is constrained to take values in a discretization S-N of the unit circle S-1 consisting o ...
2021Graph-based image representation learning
Though deep learning (DL) algorithms are very powerful for image processing tasks, they generally require a lot of data to reach their full potential. Furthermore, there is no straightforward way to impose various properties, given by the prior knowledge a ...
EPFL2019