| --- |
| license_link: https://github.com/adarshbhandaryp/ABMIL |
| datasets: |
| - ODELIA-AI/ODELIA-Challenge-2025 |
| language: |
| - en |
| metrics: |
| - roc_auc |
| pipeline_tag: image-classification |
| tags: |
| - breast |
| - cancer |
| - mri |
| - odelia |
| - abmil |
| - swin-transformer |
| extra_gated_prompt: >- |
| ### Model Usage Agreement |
| |
| By accessing or using this model (the "Model"), you acknowledge and agree |
| to the following terms: |
|
|
| 1. The Model is provided for non-commercial academic and research use only. |
| 2. The Model must not be used for diagnosis, treatment, clinical decision |
| making, or any application involving real-patient care. |
| 3. Do not identify, re-identify, or deanonymize people whose data may have |
| contributed to the training or evaluation data. |
| 4. Publications or derivative research must acknowledge the ODELIA |
| consortium and cite the ABMIL authors and source repository. |
| 5. You are responsible for validating outputs in your research context. |
| The Model is provided as-is, without warranties. |
|
|
| This gated access does not grant rights beyond the license and permissions |
| stated in this model card. Third-party components retain their own licenses. |
| --- |
| |
| # ODELIA ABMIL Model |
|
|
| Attention-Based Multiple Instance Learning (ABMIL) model for subject-level |
| three-class breast MRI classification in the ODELIA Challenge 2025. |
|
|
| This model was contributed by the ABMIL team. The source implementation is |
| available at [adarshbhandaryp/ABMIL](https://github.com/adarshbhandaryp/ABMIL). |
|
|
| ## Method |
|
|
| The model represents each subject using 32 axial slices with pre-contrast, |
| early post-contrast, and subtraction channels. A Swin Transformer Tiny |
| backbone extracts slice features. Gated attention aggregates the slice |
| features into a patient-level representation, followed by a three-class |
| classifier. The source README reports focal loss with gamma 5.0. |
|
|
| ## Intended use and limitations |
|
|
| This release is for research and benchmarking only. It is not a medical |
| device, is not clinically validated, and must not be used for diagnosis, |
| prognosis, treatment, triage, or other clinical decisions. Performance may |
| vary across institutions, scanners, acquisition protocols, and populations. |
|
|
| ## Setup |
|
|
| The source repository provides the complete training and evaluation code and |
| requirements: |
|
|
| ```bash |
| git clone https://github.com/adarshbhandaryp/ABMIL.git |
| cd ABMIL |
| pip install -r requirements.txt |
| ``` |
|
|
| Refer to the source repository for configuration-file usage and preprocessing |
| requirements. The model expects ODELIA-style multi-sequence breast MRI input. |
|
|
| ## License and attribution |
|
|
| This repository is published under the ODELIA and ABMIL authors' permission. |
|
|
| The Swin Transformer implementation and pretrained ImageNet weights are |
| third-party components. Their applicable licenses and terms must be preserved |
| and reviewed before redistribution. See the dependency notices in the source |
| repository and the relevant upstream package/model licenses. |
|
|
| Please cite or acknowledge: |
|
|
| - ODELIA Challenge 2025 and the ODELIA consortium. |
| - The ABMIL authors and [source repository](https://github.com/adarshbhandaryp/ABMIL). |
| - The ODELIA paper: [arXiv:2506.00474](https://arxiv.org/abs/2506.00474). |
|
|
