A convolutional neural network-based framework for quality control through speckle displacement analysis
Sabahno, Hamed LU
and Khodadad, Davood
(2025)
In Scientific Reports
15.
- Abstract
- Among the most advanced techniques for quality control, image processing and optical methods areprominent because of their precision and versatility. These methods often involve analyzing specklesgenerated by coherent laser illumination because coherent light provides detailed and accuratemeasurement capabilities. In speckle metrology-based techniques, the accurate measurement ofspeckle displacements is crucial for detecting faults or deformations in objects. In this study, anadvanced algorithm segments the image into overlapping grids, followed by a Fourier-based imageregistration to accurately quantify the speckle displacements. This method can simultaneously detectmultiple translational movements in the different parts of an object.... (More)
- Among the most advanced techniques for quality control, image processing and optical methods areprominent because of their precision and versatility. These methods often involve analyzing specklesgenerated by coherent laser illumination because coherent light provides detailed and accuratemeasurement capabilities. In speckle metrology-based techniques, the accurate measurement ofspeckle displacements is crucial for detecting faults or deformations in objects. In this study, anadvanced algorithm segments the image into overlapping grids, followed by a Fourier-based imageregistration to accurately quantify the speckle displacements. This method can simultaneously detectmultiple translational movements in the different parts of an object. However, proper calculationand assignment of overlap sizes to each grid plays a crucial role in this method, which is where weobtain help from convolutional neural networks (CNNs). We develop a CNN architecture and optimizeits hyperparameters using a Monte Carlo simulation algorithm incorporating a grid search and k-fold cross-validation. Finally, we validate the developed method through a case study involving asimulation and real speckle patterns generated by spraying water on a cardboard surface. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/6eb2719e-1413-41e0-8678-35000ee68ae5
- author
- Sabahno, Hamed
LU
and Khodadad, Davood
- publishing date
- 2025
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Scientific Reports
- volume
- 15
- article number
- 39192
- publisher
- Nature Publishing Group
- ISSN
- 2045-2322
- language
- English
- LU publication?
- no
- id
- 6eb2719e-1413-41e0-8678-35000ee68ae5
- date added to LUP
- 2025-11-10 14:26:32
- date last changed
- 2025-11-10 15:10:42
@article{6eb2719e-1413-41e0-8678-35000ee68ae5,
abstract = {{Among the most advanced techniques for quality control, image processing and optical methods areprominent because of their precision and versatility. These methods often involve analyzing specklesgenerated by coherent laser illumination because coherent light provides detailed and accuratemeasurement capabilities. In speckle metrology-based techniques, the accurate measurement ofspeckle displacements is crucial for detecting faults or deformations in objects. In this study, anadvanced algorithm segments the image into overlapping grids, followed by a Fourier-based imageregistration to accurately quantify the speckle displacements. This method can simultaneously detectmultiple translational movements in the different parts of an object. However, proper calculationand assignment of overlap sizes to each grid plays a crucial role in this method, which is where weobtain help from convolutional neural networks (CNNs). We develop a CNN architecture and optimizeits hyperparameters using a Monte Carlo simulation algorithm incorporating a grid search and k-fold cross-validation. Finally, we validate the developed method through a case study involving asimulation and real speckle patterns generated by spraying water on a cardboard surface.}},
author = {{Sabahno, Hamed and Khodadad, Davood}},
issn = {{2045-2322}},
language = {{eng}},
publisher = {{Nature Publishing Group}},
series = {{Scientific Reports}},
title = {{A convolutional neural network-based framework for quality control through speckle displacement analysis}},
url = {{https://lup.lub.lu.se/search/files/232639743/Sabahno_et_al-2025-Scientific_Reports.pdf}},
volume = {{15}},
year = {{2025}},
}