EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Error‐Tolerant Multimodal Vision‐Language Models for Endodontic Triaging: A Cross‐Sectional Study

Md Fahim Shahoriar Titu, Mahir Afser Pavel, Afifa Zain Apurba, Saif Ahmed iD, Shafin Rahman iD, James Dudley iD, Taseef Hasan Farook iD, Giuseppe Troiano

International Journal of Dentistry · 2026-01 · 2026 (1)

DOI: 10.1155/ijod/4148741 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Dentistry, online erschienen am 31.01.2026. Verfasst von Md Fahim Shahoriar Titu und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical and Computer Engineering, North South University, Dhaka, Bangladesh, northsouth.edu. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

To assess the performance of multimodal vision‐language artificial intelligence models, optimised using quantisation‐aware training, in triaging endodontic treatment needs. The focus is on the ability to interpret endodontic radiographs while tolerating common image capture errors, including cone cutting, elongation, foreshortening, horizontal misalignment, over‐ and under‐exposure and artefacts.

Methods

In total 3600 dental radiographs were obtained across a 1‐year period. Image augmentation techniques were applied to enhance model generalisability. Bootstrapped Language‐Image Pretraining (BLIP), CLIP, Florence 2 and Paligemma multimodal models were fine‐tuned using quantisation‐aware training and evaluated using Bilingual Evaluation Understudy (BLEU), Recall‐Oriented Understudy for Gisting Evaluation (ROUGE), Metric for Evaluation of Translation with Explicit ORdering (METEOR), Consensus‐based Image Description Evaluation (CIDEr) and Loss Trends and Convergence.

Results

Quantisation‐aware optimisation improved BLEU‐4 by at least 17.3%, METEOR by at least 11.1%, ROUGE‐L by at least 9.8% and CIDEr by at least 75.5% across all models. Quantisation reduced memory consumption by 87.5% while preserving diagnostic accuracy within a 0.5% error margin while correctly reproducing over 90% endodontic triage assessments made by practitioners.

Conclusion

Multimodal AI demonstrates tolerance to imaging inconsistencies and is capable of accurately triaging endodontic cases with minimal computational demands, without compromising diagnostic performance.

Schlagwörter

Stichwörter der Autor:innen

intraoral radiographslinguistic metricsorthopantomogramquantisation-aware trainingroot canal treatment

Autor:innen und Institutionen

  • Department of Electrical and Computer Engineering, North South University, Dhaka, Bangladesh, northsouth.edu.
  • Adelaide Dental School, University of Adelaide, School of Dentistry, Adelaide University, Adelaide, 5000, South Australia, Australia, adelaide.edu.au.

ORCID-Profile: Saif Ahmed, Shafin Rahman, James Dudley, Taseef Hasan Farook

Zitieren

APA

Titu, M. F. S., Pavel, M. A., Apurba, A. Z., Ahmed, S., Rahman, S., Dudley, J., Farook, T. H., & Troiano, G. (2026). Error‐Tolerant Multimodal Vision‐Language Models for Endodontic Triaging: A Cross‐Sectional Study. International Journal of Dentistry, 2026(1). https://doi.org/10.1155/ijod/4148741

Vancouver

Titu MFS, Pavel MA, Apurba AZ, Ahmed S, Rahman S, Dudley J, et al. Error‐Tolerant Multimodal Vision‐Language Models for Endodontic Triaging: A Cross‐Sectional Study. International Journal of Dentistry. 2026;2026(1). doi:10.1155/ijod/4148741

BibTeX
@article{Titu2026,
  author = {Titu, Md Fahim Shahoriar and Pavel, Mahir Afser and Apurba, Afifa Zain and Ahmed, Saif and Rahman, Shafin and Dudley, James and Farook, Taseef Hasan and Troiano, Giuseppe},
  title = {Error‐Tolerant Multimodal Vision‐Language Models for Endodontic Triaging: A Cross‐Sectional Study},
  journal = {International Journal of Dentistry},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/ijod/4148741}
}
Kostenlos starten →