Views
No views yet
| Model | Description |
|---|---|
baseline-vit-transformer | ViT frame encoder with transformer-based temporal aggregation |
maskloss-block7-3heads | Mask-guided attention loss applied to three heads at code block index 7 |
maskloss-block11-3heads | Mask-guided attention loss applied to three heads at code block index 11 |
maskloss-block11-6heads | Mask-guided attention loss applied to six heads at code block index 11 |
7 and 11 correspond to Blocks 8 and 12 in the paper..zip archive containing checkpoints and training outputs for all five patient-level cross-validation folds.1model-name/
2├── config.json
3├── final_metrics.json
4├── fold_0/
5│ ├── best.pt
6│ └── best_val_metrics.json
7├── fold_1/
8├── fold_2/
9├── fold_3/
10└── fold_4/best.pt as the selected checkpoint for each fold.1@article{almsouti2026hierarchy,
2 title = {Hierarchy-Aware and Anatomy-Guided Learning for Lung Ultrasound Video Classification},
3 author = {Almsouti, Alya and Mecharbat, Lotfi and Aboukhater, Noha and Alabrach, Yousef and Anwar, Siddiq and Kumar, Andre and Almakky, Ibrahim and Yaqub, Mohammad},
4 journal = {arXiv preprint arXiv:2607.17551},
5 year = {2026},
6 doi = {10.48550/arXiv.2607.17551}
7}