CXR-LT 2026 Challenge: Projection-Aware Multi-Label and Zero-Shot Chest X-Ray Classification
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914440371240960 |
|---|---|
| author | Cho, Juno Kim, Dohui Kim, Mingeon Jang, Hyunseo Lee, Chang Sun Ye, Jong Chul |
| author_facet | Cho, Juno Kim, Dohui Kim, Mingeon Jang, Hyunseo Lee, Chang Sun Ye, Jong Chul |
| contents | This challenge tackles multi-label classification for known chest X-ray (CXR) lesions and zero-shot classification for unseen ones. To handle diverse CXR projections, we integrate projection-specific models via a classification network into a unified framework. For zero-shot classification (Task 2), we extend CheXzero with a novel dual-branch architecture that combines contrastive learning, Asymmetric Loss (ASL), and LLM-generated descriptive prompts. This effectively mitigates severe long-tail imbalances and maximizes zero-shot generalization. Additionally, strong data and test-time augmentations (TTA) ensure robustness across both tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_02185 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CXR-LT 2026 Challenge: Projection-Aware Multi-Label and Zero-Shot Chest X-Ray Classification Cho, Juno Kim, Dohui Kim, Mingeon Jang, Hyunseo Lee, Chang Sun Ye, Jong Chul Computer Vision and Pattern Recognition This challenge tackles multi-label classification for known chest X-ray (CXR) lesions and zero-shot classification for unseen ones. To handle diverse CXR projections, we integrate projection-specific models via a classification network into a unified framework. For zero-shot classification (Task 2), we extend CheXzero with a novel dual-branch architecture that combines contrastive learning, Asymmetric Loss (ASL), and LLM-generated descriptive prompts. This effectively mitigates severe long-tail imbalances and maximizes zero-shot generalization. Additionally, strong data and test-time augmentations (TTA) ensure robustness across both tasks. |
| title | CXR-LT 2026 Challenge: Projection-Aware Multi-Label and Zero-Shot Chest X-Ray Classification |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.02185 |