Multiple approaches to use deep neural networks for image restoration have recently been proposed. Training such networks requires well registered pairs of high and low-quality images. While this is easily achievable for many imaging modalities, e.g., fluorescence light microscopy, for others it is not. Here we summarize on a number of recent developments in the fast-paced field of Content-Aware Image Restoration (CARE), in particular, and the associated area of neural network training, more in general. We then give specific examples how electron microscopy data can benefit from these new technologies.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1016/bs.mcb.2019.05.001 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Intervertebral disc regeneration - Is it possible?
Acta Orthop Traumatol Turc
December 2024
Department of Orthopedics and Traumatology, Brugmann University Hospital Center, Free University of Brussels, Brussels, Belgium.
Objective: The aim of this study was to evaluate disc metabolism after decreasing the axial load through surgery by assessing the glycosaminoglycan content through a non-invasive method-delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC).
Methods: Sixteen patients with mono-segmental disc degeneration (L4-L5 or L5-S1) who underwent posterior lumbar spine fixation with intervertebral distraction of 2 consecutive vertebrae using monoaxial transpedicular screws and lyophilized allograft to achieve segmental fusion, and who had a follow-up period of at least 2 years, were included in this study. The first lumbar disc was used as the control group.
Underwater optical imaging, especially in coastal waters, suffers from reduced spatial resolution and contrast by forward scattered light. With the increased number of hyper- and multi-spectral imaging applications, the effect of the point spread function (PSF) at different spectral bands becomes increasingly more relevant. In this work, extensive laboratory measurements of the PSF at 450, 500, 550, 600 and 650 nm in different turbidity have been carried out.
View Article and Find Full Text PDFThis paper proposes an imaging technique to remove strong reflections from specular surfaces using polarization characteristics combined with light field imaging. Firstly, the correct strong reflection region is found by studying the reflected light characteristics, and the strong reflection region highlights are filtered out using Stokes parameters based on polarization information. Then, a system of microlens arrays with different transmittances was built for imaging, and the system was image-corrected to enable more information about the scene to be captured.
View Article and Find Full Text PDFAtmospheric turbulence introduces random disturbances that degrade and distort images of observed targets as light propagates through the atmosphere. Although numerous algorithms have been developed to restore images degraded by turbulence, most of these algorithms lack sufficient generalization and are limited to specific application scenarios or fixed atmospheric turbulence intensities. In this paper, we propose an Atmospheric Turbulence Restoration Network (ATRN), a two-stage algorithm based on multi-frame information fusion.
View Article and Find Full Text PDFOur study introduces a pioneering underwater single-pixel imaging approach that employs an orbital angular momentum (OAM) basis as a sampling scheme and a dual-attention residual U-Net generative adversarial network (DARU-GAN) as reconstruction algorithm. This method is designed to address the challenges of low sampling rates and high turbidity typically encountered in underwater environments. The integration of the OAM-basis sampling scheme and the improved reconstruction network not only enhances reconstruction quality but also ensures robust generalization capabilities, effectively restoring underwater target images even under the stringent conditions of a 3.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!