Federico Bolelli
AI4M3D Workshop · ECCV 2026 · Invited Talk

From Volumes to Clinical Decision

Five conditions to fill the gap between medical 3D vision and the clinic

Shared Data · Reliable Supervision · Robust Models · Actionable Outputs · Trusted Evaluation

Federico Bolelli — AImageLab, University of Modena and Reggio Emilia
federico.bolelli@unimore.it · September 9, 2026

Standardized and Shared Data

Shared cohorts, controlled evaluation and standardized evidence — the ToothFairy challenge series and what it taught us about benchmarking.

Segmenting the Inferior Alveolar Canal in CBCT Volumes: the ToothFairy Challenge

Federico Bolelli, Luca Lumetti, Shankeeth Vinayahalingam, ..., Alexandre Anesi, Costantino Grana

IEEE Transactions on Medical Imaging, Dec 2024

The first ToothFairy challenge: inferior alveolar canal segmentation in CBCT.

Multi-Structure Segmentation in CBCT Volumes: the ToothFairy2 Challenge

Federico Bolelli, Luca Lumetti, Niels van Nistelrooij, Shankeeth Vinayahalingam, Mattia Di Bartolomeo, Kevin Marchesini, Arrigo Pellacani, Ettore Candeloro, Gabriele Rosati, Tong Xi, Fabian Isensee, Yannick Kirchhoff, Lars Krämer, Maximilian Rokuss, Constantin Ulrich, Klaus Maier-Hein, Yuxian Jiang, Yusheng Liu, Lisheng Wang, Haoshen Wang, Siyu Chen, Zhiming Cui, Pengcheng Shi, Zhaohong Pan, Xiaokun Liang, Qi Ma, Ender Konukoglu, Marek Wodzinski, Henning Müller, Haipeng Mai, Xiaobing Dang, Shrajan Bhandary, Radu Grosu, Stefaan Bergé, Alexandre Anesi, Costantino Grana

Medical Image Analysis, April 2026

ToothFairy2: multi-structure segmentation, 42 classes.

ToothFairy3: Scaling CBCT Maxillofacial Segmentation to 77 Classes with U-Mamba2

Luca Lumetti, Zhi Qin Tan, Lorenzo Borghi, Niels van Nistelrooij, Gabriele Rosati, Owen Addison, Yupeng Li, Shankeeth Vinayahalingam, Costantino Grana, Federico Bolelli

Medical Image Computing and Computer Assisted Intervention – MICCAI 2026, May 2026

ToothFairy3: scaling to 77 classes with U-Mamba2.

Segmenting Maxillofacial Structures in CBCT Volumes

Federico Bolelli, Kevin Marchesini, Niels van Nistelrooij, Luca Lumetti, Vittorio Pipoli, Elisa Ficarra, Shankeeth Vinayahalingam, Costantino Grana

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Mar 2025

The nnU-Net lesson: once the benchmark is controlled, the pipeline matters as much as the backbone.

ToothFairy on Grand Challenge

Challenge portal — ToothFairy, ToothFairy2 and ToothFairy3 (ToothFairy4 runs as Task 1 of ODIN 2026)

ToothFairy toolkit and baselines

Code, evaluation scripts and baselines for the whole series

Ditto — the AImageLab dataset platform

Where our datasets are versioned, documented and shared

Reliable Supervision

Dense 3D masks are the exception; sparse labels are what clinics already produce. Propagating them, and encoding the annotation protocol in the tooling.

Improving Segmentation of the Inferior Alveolar Nerve through Deep Label Propagation

Marco Cipriano, Stefano Allegretti, Federico Bolelli, Federico Pollastri, Costantino Grana

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2022

Deep label propagation: turning sparse annotations into dense supervision.

Do Multimodal LLMs Understand Intraoral Dental Data? Dataset, Platform, and Baselines

Luca Lumetti, Federico Rizzo, Francesca Cremonini, Ettore Candeloro, Luca Lombardo, Costantino Grana, Federico Bolelli

European Conference on Computer Vision (ECCV), June 2026

Dataset, platform and baselines for intraoral dental data — the work behind Yggdrasil.

Yggdrasil — collaborative annotation platform

Patient-centric ingestion, joint multi-modal visualization, protocol encoded in the platform

Well-defined Input and Tasks

Missing modalities, domain shift and a task that keeps changing after deployment — three ways the input assumption breaks in practice.

IM-Fuse: A Mamba-based Fusion Block for Brain Tumor Segmentation with Incomplete Modalities

Vittorio Pipoli, Alessia Saporita, Kevin Marchesini, Costantino Grana, Elisa Ficarra, Federico Bolelli

Medical Image Computing and Computer Assisted Intervention – MICCAI 2025, May 2025

IM-Fuse: graceful degradation when MRI sequences are missing.

A New Multicenter Testicular US Dataset and a Lightweight Cond-UNet for Generalization in US Segmentation

Nicola Morelli, Kevin Marchesini, Daniele Santi, Costantino Grana, Federico Bolelli

Proceedings of the British Machine Vision Conference, August 2026

Cond-UNet: a lightweight conditioned U-Net that beats a foundation model on multicenter testicular ultrasound.

U-Net Transplant: The Role of Pre-training for Model Merging in 3D Medical Segmentation

Luca Lumetti, Giacomo Capitani, Elisa Ficarra, Costantino Grana, Simone Calderara, Angelo Porrello, Federico Bolelli

Medical Image Computing and Computer Assisted Intervention – MICCAI 2025, May 2025

U-Net Transplant: add a structure by merging specialists instead of retraining.

Actionable Output

Segmentation describes anatomy; reports communicate clinical meaning. ToothFairy4 moves the benchmark from masks to structured findings.

Ontology-Grounded Structured Prediction for Dental CBCT Reporting

Luca Lumetti, Mattia Di Bartolomeo, Arrigo Pellacani, Alexandre Anesi, Costantino Grana, Federico Bolelli

Medical Image Computing and Computer Assisted Intervention – MICCAI 2026, June 2026

Ontology-grounded structured prediction: 893 reports over 529 public CBCT volumes, a clinician-designed OWL ontology and six stage-aware metrics.

ToothFairy4 — Maxillofacial and Surgical Report Generation from CBCT

Task 1 of the ODIN 2026 challenge on Grand Challenge — the benchmark moves from masks to structured findings

Trusted Evaluation

Lexical metrics score how a report is phrased, not whether it is true — and a published Dice does not tell a hospital which model to deploy.

Unsupervised Source-Free Ranking of Biomedical Segmentation Models Under Distribution Shift

Joshua Talks, Kevin Marchesini, Luca Lumetti, Federico Bolelli, Anna Kreshuk

European Conference on Computer Vision (ECCV), June 2026

Ranking segmentation models with no labels, no source data and no weights — only the predictions.

Our Research Group

None of this is single-author work. Code for the papers above lives in our group GitHub organization.

AImageLab-zip on GitHub

Our group GitHub organization — code for the works above and more

AImageLab

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.