Three-Dimensional Thermal Tomography with Physics-Informed Neural Networks
Tomography · 30.11.2024 · 10 (12) · 1930-1946
DOI: 10.3390/tomography10120140 ↗Kurz gefasst
Fachartikel in der Zeitschrift Tomography, online erschienen am 30.11.2024. Verfasst von Theodoros Leontiou und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Mechanical Engineering, Frederick University, Nicosia 1036, Cyprus. Hauptthemen laut MEDLINE: Neural Networks, Computer, Phantoms, Imaging, Imaging, Three-Dimensional, Thermography. Die Arbeit wurde bisher 10-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research and Innovation Foundation, Cyprus Research and Innovation Foundation.
Abstract (Originalsprache)
Background
Accurate reconstruction of internal temperature fields from surface temperature data is critical for applications such as non-invasive thermal imaging, particularly in scenarios involving small temperature gradients, like those in the human body.
Methods
In this study, we employed 3D convolutional neural networks (CNNs) to predict internal temperature fields. The network’s performance was evaluated under both ideal and non-ideal conditions, incorporating noise and background temperature variations. A physics-informed loss function embedding the heat equation was used in conjunction with statistical uncertainty during training to simulate realistic scenarios.
Results
The CNN achieved high accuracy for small phantoms (e.g., 10 cm in diameter). However, under non-ideal conditions, the network’s predictive capacity diminished in larger domains, particularly in regions distant from the surface. The introduction of physical constraints in the training processes improved the model’s robustness in noisy environments, enabling accurate reconstruction of hot-spots in deeper regions where traditional CNNs struggled.
Conclusions
Combining deep learning with physical constraints provides a robust framework for non-invasive thermal imaging and other applications requiring high-precision temperature field reconstruction, particularly under non-ideal conditions.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Mechanical Engineering, Frederick University, Nicosia 1036, Cyprus.
- Computation-Based Science and Technology Research Center (CaSToRC), The Cyprus Institute, 20 Kavafi Street, Nicosia 2121, Cyprus.
- Department of Business Administration, Frederick University, Nicosia 1036, Cyprus.
- Department of Physics, National and Kapodistrian University of Athens, Zografou, GR-157 84 Athens, Greece.
- KYAMOS Ltd., 37 Polyneikis Street, Strovolos, Nicosia 2047, Cyprus.
- Department of Electrical, Computer Engineering and Informatics, Frederick University, Nicosia 1036, Cyprus.
ORCID-Profile: Theodoros Leontiou, Marios Charalambides, Efstathios Stiliaris, Costas N Papanicolas
Förderung
- Research and Innovation Foundation
- Cyprus Research and Innovation Foundation
Zitieren
APA
Leontiou, T., Frixou, A., Charalambides, M., Stiliaris, E., Papanicolas, C. N., Nikolaidou, S., & Papadakis, A. (2024). Three-Dimensional Thermal Tomography with Physics-Informed Neural Networks. Tomography, 10(12), 1930-1946. https://doi.org/10.3390/tomography10120140
Vancouver
Leontiou T, Frixou A, Charalambides M, Stiliaris E, Papanicolas CN, Nikolaidou S, et al. Three-Dimensional Thermal Tomography with Physics-Informed Neural Networks. Tomography. 2024;10(12):1930-1946. doi:10.3390/tomography10120140
BibTeX
@article{Leontiou2024,
author = {Leontiou, Theodoros and Frixou, Anna and Charalambides, Marios and Stiliaris, Efstathios and Papanicolas, Costas N and Nikolaidou, Sofia and Papadakis, Antonis},
title = {Three-Dimensional Thermal Tomography with Physics-Informed Neural Networks},
journal = {Tomography},
year = {2024},
volume = {10},
number = {12},
pages = {1930--1946},
doi = {10.3390/tomography10120140}
}
EUVIMED