Research map / Label-efficient and robust learning / Self-supervised and contrastive
Self-Supervised Pre-Training of Swin Transformers for 3D Medical Image Analysis
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Pre-train Swin UNETR with unlabeled CT images and fine-tune for segmentation (ablation) | Swin UNETR · BTCV | Dice: 83.43 -> 84.72 | Helped |
| Pre-train Swin UNETR with unlabeled CT images and fine-tune for segmentation (ablation) | Swin UNETR · BTCV | HD: 42.36 -> 20.03 | Helped |
| Add masked volume inpainting as the only pre-training objective (ablation) | Swin UNETR · BTCV | Dice: 83.43 -> 83.85 | Helped |
| Add rotation prediction as the only pre-training objective (ablation) | Swin UNETR · BTCV | Dice: 83.43 -> 83.56 | Helped |
| Add contrastive coding as the only pre-training objective (ablation) | Swin UNETR · BTCV | Dice: 83.43 -> 83.67 | Helped |
| Combine inpainting and rotation prediction, without contrastive coding (ablation) | Swin UNETR · BTCV | Dice: 84.01 | Helped |
| Combine inpainting and contrastive coding, without rotation prediction (ablation) | Swin UNETR · BTCV | Dice: 84.45 | Helped |
| Add rotation prediction to inpainting and contrastive coding (ablation) | Swin UNETR · BTCV | Dice: 84.45 -> 84.72 | Helped |
| Add rotation prediction to inpainting and contrastive coding (ablation) | Swin UNETR · BTCV | HD: 24.37 -> 20.03 | Helped |
| Use self-supervised pre-training before fine-tuning (ablation) | Swin UNETR · MSD Task03 Liver | Dice: 75.27% -> 77.77% | Helped |
| Use self-supervised pre-training before fine-tuning (ablation) | Swin UNETR · MSD Task08 Hepatic Vessel | Dice: 64.63% -> 68.52% | Helped |
| Use self-supervised pre-training before fine-tuning (ablation) | Swin UNETR · MSD Task10 Colon | Dice: 34.83% -> 43.38% | Helped |
Closest papers by meaning
- Revisiting MAE Pre-training for 3D Medical Image Segmentation (CVPR 2025 · similarity 0.89)
- nnWNet: Rethinking the Use of Transformers in Biomedical Image Segmentation and Calling for a Unified Evaluation Benchmark (CVPR 2025 · similarity 0.87)
- How Well Do Supervised 3D Models Transfer to Medical Imaging Tasks? (ICLR 2024 · similarity 0.86)
- DDViT: Double-Level Fusion Domain Adapter Vision Transformer (Student Abstract) (AAAI 2024 · similarity 0.85)
- Autoregressive Sequence Modeling for 3D Medical Image Representation (AAAI 2025 · similarity 0.84)
- Time-to-Event Pretraining for 3D Medical Imaging (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.