Journal of Modern Research Physics
Volume 10, Issue 2 (Autumn and Winter 2026)
JMRPh 2026, 10(2): 26-37 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Yaghoobi M, Jafarfard M R, Zare B. Speckle noise elimination of phase images using convolution neural networks. JMRPh 2026; 10 (2) :26-37
URL: http://jmrph.khu.ac.ir/article-1-281-en.html
URL: http://jmrph.khu.ac.ir/article-1-281-en.html
Abstract: (198 Views)
Phase images obtained from holograms and interference patterns in quantitative phase microscopy often suffer from distortions, phase wrapping, and phase noise. Despite extensive efforts over the past decades and the development of various denoising methods, these challenges have not yet been completely resolved. Classical noise reduction filters typically damage image details, reduce overall quality, and weaken boundary and edge detection, thereby eliminating useful image information. A promising modern approach involves the use of machine learning algorithms. Results from image quality assessment metrics show that convolutional neural networks outperform conventional phase image denoising methods. In this approach, training data are fed into the network, and after several training stages, the network acquires the ability to reduce noise in phase images with high accuracy.
Type of Study: Research |
Received: 2025/12/17 | Accepted: 2026/02/9 | Published: 2026/03/20 | ePublished: 2026/03/20
Received: 2025/12/17 | Accepted: 2026/02/9 | Published: 2026/03/20 | ePublished: 2026/03/20
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |