Image restoration in frequency space using complex-valued CNNs

Real-valued convolutional neural networks (RV-CNNs) in the spatial domain have outperformed classical approaches in many image restoration tasks such as image denoising and super-resolution. Fourier analysis of the results produced by these spatial domain models reveals the limitations of these mode...

Full description

Bibliographic Details
Published in:Frontiers in Artificial Intelligence
Main Authors: Zafran Hussain Shah, Marcel Müller, Wolfgang Hübner, Henning Ortkrass, Barbara Hammer, Thomas Huser, Wolfram Schenck
Format: Article
Language:English
Published: Frontiers Media S.A. 2024-09-01
Subjects:
image restoration
image denoising
super-resolution
convolutional neural networks (CNNs)
complex-valued convolutional neural networks (CV-CNNs)
complex-valued attention gates
Online Access:https://www.frontiersin.org/articles/10.3389/frai.2024.1353873/full

Internet

https://www.frontiersin.org/articles/10.3389/frai.2024.1353873/full

Similar Items