Dati, paradigmi di apprendimento e ricerca applicata all’imaging dentale
AI in Medical Imaging · How AI Learns · What We Do · Take-Home
Many teams, one clinical task, the same open data and a sealed answer key: four editions of ToothFairy, from one nerve canal to 77 structures and clinical reports.
IEEE Transactions on Medical Imaging, Dec 2024
The first ToothFairy challenge: inferior alveolar canal segmentation in CBCT.
Medical Image Analysis, April 2026
ToothFairy2: multi-structure segmentation, 42 classes.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2026, May 2026
ToothFairy3: 77 classes per scan, including the 32 dental pulps.
Challenge portal for the whole series
Code, evaluation scripts and baselines for the whole series
Every annotation costs specialist time. Dense masks drawn from scratch, sparse labels the clinic already produces, and a few clicks to correct what the AI proposes.
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Mar 2025
Dense annotation from scratch: maxillofacial structures in CBCT.
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2022
Deep label propagation: routine sparse labels turned into dense 3D masks.
Structured Challenge, Sep 2025
ToothFairy3 interactive track: the clinician clicks, the AI corrects the mask.
Generative models learn from real images, then create patients who never existed: for sharing, rare cases and more training data, always validated on real patients.
Image Analysis and Processing – ICIAP 2025, Sep 2025
Synthetic testicular ultrasound: the "real or synthetic?" images from the talk.
2018 IEEE 31st International Symposium on Computer-Based Medical Systems (CBMS 2018), Jun 2018
Synthetic skin lesions generated together with their masks.
BraTS 2025 Lighthouse Challenge, MICCAI Workshop, Sep 2025
Synthesizing missing brain MRI sequences for the BraSyn challenge.
One patient, many modalities: CBCT, intraoral scans and photographs, and the reports orthodontists write about them.
European Conference on Computer Vision (ECCV), June 2026
Bite2Text: 1,000 orthodontic patients, and an AI that writes the report from the intraoral scan.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2026, June 2026
From segmentation to structured CBCT reports (ToothFairy4).
Proceedings of the British Machine Vision Conference, Sep 2026
The same idea in pathology: the report as guideline-aligned classification instead of free text.
Oral and Dental Image Analysis - ODIN, MICCAI Workshop, Sep 2025
Bits2Bites: occlusal classification straight from the intraoral scan.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2026 Satellite Events, August 2026
IOP-Compass: tooth segmentation with FDI numbering in intraoral photographs.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2026 Satellite Events, August 2026
Landmark-guided registration of intraoral scan and CBCT.
Data organized around the patient, modalities viewed together, the annotation protocol encoded in the platform
AI in orthodontics will be as good as the data and the evaluation you help build.
Blind comparison of orthodontic reports, two minutes per case
Open data: ToothFairy, Bite2Text, IOP-Compass and more
None of this is single-author work. Code for the papers above lives in our group GitHub organization.
Our group GitHub organization — code for the works above and more
AImageLab, University of Modena and Reggio Emilia
Something missing or a broken link? Let me know. A complete list of publications is available on the publications page.