Research map / Detection and segmentation / Open-set and few-shot detection
MaskSAM: Auto-prompt SAM with Mask Classification for Volumetric Medical Image Segmentation
Bin Xie, Hao Tang, Bin Duan, Dawen Cai, Yan Yan, Gady Agam
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use averaged mask-derived and learnable box prompts instead of masks only (ablation) | MaskSAM · ACDC | Dice (DSC, %): 89.53 -> 91.45 | Improved |
| use averaged mask-derived and learnable box prompts instead of boxes only (ablation) | MaskSAM · ACDC | Dice (DSC, %): 88.78 -> 91.45 | Improved |
| use averaged boxes instead of separate mask and box prompts (ablation) | MaskSAM · ACDC | Dice (DSC, %): 90.08 -> 91.45 | Improved |
| add depth positional embedding (ablation) | MaskSAM · ACDC | Dice (DSC, %): 91.45 -> 91.61 | Improved |
| add depth MLP after FC-Up of vanilla adapter (ablation) | MaskSAM · ACDC | Dice (DSC, %): 91.61 -> 92.88 | Improved |
| add depth MLP before FC-Down of vanilla adapter (ablation) | MaskSAM · ACDC | Dice (DSC, %): 91.61 -> 92.93 | Improved |
| use DMLPAdapter (depth MLP in middle of adapter) (ablation) | MaskSAM · ACDC | Dice (DSC, %): 91.61 -> 93.10 | Improved |
| add DConvAdapter for image embeddings on top of DMLPAdapter (ablation) | MaskSAM · ACDC | Dice (DSC, %): 93.10 -> 93.39 | Improved |
| fine-tune SAM with MaskSAM instead of SAMed (prompt-free) | SAM · AMOS22 test | Dice (DSC, %): 0.772 -> 0.905 | Improved |
| use nnUNet-predicted box prompts with SAM instead of nnUNet itself | SAM · AMOS22 test | Dice (DSC, %): 0.878 -> 0.538 | Hurt |
| replace nnUNet with MaskSAM | nnUNet · Synapse | Dice (DSC, %): 86.21 -> 87.23 | Improved |
| replace nnUNet with MaskSAM | nnUNet · ACDC | Dice (DSC, %): 91.61 -> 93.39 | Improved |
Closest papers by meaning
- Unleashing the Potential of SAM for Medical Adaptation via Hierarchical Decoding (CVPR 2024 · similarity 0.90)
- Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.89)
- AoP-SAM: Automation of Prompts for Efficient Segmentation (AAAI 2025 · similarity 0.89)
- BLO-SAM: Bi-level Optimization Based Finetuning of the Segment Anything Model for Overfitting-Preventing Semantic Segmentation (ICML 2024 · similarity 0.88)
- SAMRefiner: Taming Segment Anything Model for Universal Mask Refinement (ICLR 2025 · similarity 0.87)
- Improving the Generalization of Segmentation Foundation Model under Distribution Shift via Weakly Supervised Adaptation (CVPR 2024 · similarity 0.87)
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.
