IJAIN Reading Tools

Privacy-Preserving U-Net Variants with pseudo-labeling for radiolucent lesion segmentation in dental CBCT

Dublin Core		PKP Metadata Items	Metadata for this Document
1.	Title	Title of document	Privacy-Preserving U-Net Variants with pseudo-labeling for radiolucent lesion segmentation in dental CBCT
2.	Creator	Author's name, affiliation, country	Amelia Ritahani Ismail; Department of Computer Science, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Faris Farhan Azlan; Department of Computer Science, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Khairul Akmal Noormaizan; Department of Computer Science, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Nurul Afiqa; Department of Computer Science, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Syed Qamrun Nisa; Department of Computer Science, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Ahmad Badaruddin Ghazali; Department of Oral Maxillofacial Surgery & Oral Diagnosis, International Islamic University Malaysia; Malaysia
2.	Creator	Author's name, affiliation, country	Andri Pranolo; Universitas Ahmad Dahlan; Indonesia
2.	Creator	Author's name, affiliation, country	Shoffan Saifullah; AGH University of Science and Technology; Poland
3.	Subject	Discipline(s)
3.	Subject	Keyword(s)
4.	Description	Abstract	Accurate segmentation of radiolucent lesions in dental Cone-Beam Computed Tomography (CBCT) is vital for enhancing diagnostic reliability and reducing the burden on clinicians. This study proposes a privacy-preserving segmentation framework leveraging multiple U-Net variants—U-Net, DoubleU-Net, U2-Net, and Spatial Attention U-Net (SA-UNet)—to address challenges posed by limited labeled data and patient confidentiality concerns. To safeguard sensitive information, Differential Privacy Stochastic Gradient Descent (DP-SGD) is integrated using TensorFlow-Privacy, achieving a privacy budget of ε ≈ 1.5 with minimal performance degradation. Among the evaluated architectures, U2-Net demonstrates superior segmentation performance with a Dice coefficient of 0.833 and an Intersection over Union (IoU) of 0.881, showing less than 2% reduction under privacy constraints. To mitigate data annotation scarcity, a pseudo-labeling approach is implemented within an MLOps pipeline, enabling semi-supervised learning from unlabeled CBCT images. Over three iterative refinements, the pseudo-labeling strategy reduces validation loss by 14.4% and improves Dice score by 2.6%, demonstrating its effectiveness. Additionally, comparative evaluations reveal that SA-UNet offers competitive accuracy with faster inference time (22 ms per slice), making it suitable for low-resource deployments. The proposed approach presents a scalable and privacy-compliant framework for radiolucent lesion segmentation, supporting clinical decision-making in real-world dental imaging scenarios.
5.	Publisher	Organizing agency, location	Universitas Ahmad Dahlan
6.	Contributor	Sponsor(s)
7.	Date	(YYYY-MM-DD)	2025-05-31
8.	Type	Status & genre	Peer-reviewed Article
8.	Type	Type
9.	Format	File format	PDF
10.	Identifier	Uniform Resource Identifier	https://ijain.org/index.php/IJAIN/article/view/1529
10.	Identifier	Digital Object Identifier (DOI)	https://doi.org/10.26555/ijain.v11i2.1529
11.	Source	Title; vol., no. (year)	International Journal of Advances in Intelligent Informatics; Vol 11, No 2 (2025): May 2025
12.	Language	English=en	en
13.	Relation	Supp. Files
14.	Coverage	Geo-spatial location, chronological period, research sample (gender, age, etc.)
15.	Rights	Copyright and permissions	Copyright (c) 2025 Amelia Ritahani Ismail, Faris Farhan Azlan, Khairul Akmal Noormaizan, Syed Qamrun Nisa, Ahmad Badaruddin Ghazali, Andri Pranolo, Shoffan Saifullah This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.