Cross-Domain Generalization of Deep Learning Architectures for Cephalometric Landmark Detection: A Dual-Dataset and Multi-Device Benchmark
Diagnostics · 26.08.2026 · 16 (17) · 2726
DOI: 10.3390/diagnostics16172726 ↗Kurz gefasst
Fachartikel in der Zeitschrift Diagnostics, online erschienen am 26.08.2026. Verfasst von Mustafa Özcan und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Orthodontics, Faculty of Dentistry, Istanbul Health and Technology University, 34275 Istanbul, Türkiye. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background/Objectives: Deep learning models for cephalometric landmark detection report near-ceiling accuracy on single benchmarks, yet most are trained and tested on the same dataset. Whether the best in-domain architecture remains best out-of-domain has not been systematically quantified.
Methods
Four architecture families (heatmap CNN, two-stage cascade, coordinate-regression Vision Transformer, pretrained ResNet-50) were each trained on two independently sourced datasets—ISBI 2015 (400 images, one device) and Aariz (1000 images, seven devices)—and evaluated on both, over their 19 shared landmarks in millimeters. All axes used three seeds (mean ± SD): the cross-dataset matrix, leave-one-device-out shift, balanced joint training, a pretrained-versus-scratch ablation, and a landmark-level breakdown.
Results
In-domain mean radial error (MRE) was 2.47–3.04 mm (Aariz) and 4.12–5.55 mm (ISBI); cross-dataset error rose steeply (Generalization Drop—the relative increase in error out-of-domain—155–426%). The most accurate model in-domain (a from-scratch heatmap CNN, 2.47 mm) showed the largest drop, and every pretrained estimate fell below every from-scratch estimate (family means 248% vs. 380%): in-domain ranking did not predict cross-domain ranking. Leave-one-device-out revealed reproducible device-specific shift (held-out MRE 1.89–7.94 mm). Inter-observer variability was 0.53 mm, so cross-domain errors were 15–44× the human band. Balanced joint training reduced the cross-domain gap for all four architectures (both domains ≈ 2.1–3.5 mm) without harming in-domain accuracy. Pretraining more than halved cross-domain error from Aariz to ISBI (7.73 vs. 16.36 mm) but not in reverse, supporting the mechanism in one direction rather than uniformly. A-point, B-point, Nasion, and Menton all exceeded the 2 mm clinical threshold out-of-domain, including points localized to sub-millimeter accuracy in-domain.
Conclusions
Single-dataset accuracy substantially overstates clinical generalizability, and the best in-domain architecture is not the most transferable, so in-domain leaderboards are an unreliable basis for selecting a model for clinical deployment; balanced multi-source training recovers most of the loss across the architectures tested.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Orthodontics, Faculty of Dentistry, Istanbul Health and Technology University, 34275 Istanbul, Türkiye.
ORCID-Profile: Mustafa Özcan, Ferdi Allaf
Zitieren
APA
Özcan, M., & Allaf, F. (2026). Cross-Domain Generalization of Deep Learning Architectures for Cephalometric Landmark Detection: A Dual-Dataset and Multi-Device Benchmark. Diagnostics, 16(17), 2726. https://doi.org/10.3390/diagnostics16172726
Vancouver
Özcan M, Allaf F. Cross-Domain Generalization of Deep Learning Architectures for Cephalometric Landmark Detection: A Dual-Dataset and Multi-Device Benchmark. Diagnostics. 2026;16(17):2726. doi:10.3390/diagnostics16172726
BibTeX
@article{zcan2026,
author = {Özcan, Mustafa and Allaf, Ferdi},
title = {Cross-Domain Generalization of Deep Learning Architectures for Cephalometric Landmark Detection: A Dual-Dataset and Multi-Device Benchmark},
journal = {Diagnostics},
year = {2026},
volume = {16},
number = {17},
pages = {2726},
doi = {10.3390/diagnostics16172726}
}
EUVIMED