The Problem Of Image Super-Resolution, Denoising And Some Image Restoration Methods In Deep Learning Models

In this article, we address the challenges of image super-resolution and noise reduction, which are crucial for enhancing the quality of images derived from low-resolution or noisy data. We compared and assessed several approaches for upgrading low-resolution images to higher resolutions and for eli...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-06
Main Authors: Pham, Ngoc-Giau, Thanh-Hai Tong Le, Van-Hieu, Duong, Hong-Ngoc Tran, Phuoc-Hung Vo
Format: Article
Language:English
Subjects:
Deep learning
Image enhancement
Image quality
Image resolution
Image restoration
Neural networks
Noise reduction
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this article, we address the challenges of image super-resolution and noise reduction, which are crucial for enhancing the quality of images derived from low-resolution or noisy data. We compared and assessed several approaches for upgrading low-resolution images to higher resolutions and for eliminating unwanted noise, all while maintaining the essential characteristics of the original images and recovering images from poor quality or damaged data using deep learning models. Our analysis and the experimental outcomes on image quality metrics indicate that the EDCNN neural network model, enhanced with pretrained weights, significantly outperforms other methods with a Train PSNR of 31.215, a Valid PSNR of 29.493, and a Test PSNR of 31.6632.
ISSN:2331-8422