Research map / Detection and segmentation / Open-set and few-shot detection
K-Prism: A Knowledge-Guided and Prompt Integrated Universal Medical Image Segmentation Model
Bangwei Guo, Yunhe Gao, Meng Ye, Difei Gu, Yang Zhou, Leon Axel, Dimitris N. Metaxas
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 |
|---|---|---|---|
| add MoE cross-attention (ablation) | K-Prism decoder · semantic, ACDC+BUS | Dice: 77.38 -> 81.28 | Improved |
| add MoE FFN (ablation) | K-Prism decoder · semantic, ACDC+BUS | Dice: 78.57 -> 81.28 | Improved |
| add both MoE CA and FFN (ablation) | K-Prism decoder · semantic, ACDC+BUS | Dice: 76.77 -> 81.28 | Improved |
| add MoE CA and FFN (ablation) | K-Prism decoder · in-context, ACDC+BUS | Dice: 75.10 -> 79.21 | Improved |
| add MoE CA and FFN (ablation) | K-Prism decoder · interactive, ACDC+BUS | Dice(5): 92.22 -> 93.79 | Improved |
| add 2-D dense fusion (ablation) | K-Prism in-context mode · in-context, ACDC+BUS | Dice: 54.65 -> 80.84 | Improved |
| add 1-D sparse queries (ablation) | K-Prism in-context mode · in-context, ACDC+BUS | Dice: 77.19 -> 80.84 | Improved |
| add 2-D dense fusion (ablation) | K-Prism interactive mode · interactive, ACDC+BUS | Dice(5): 79.21 -> 93.68 | Improved |
| add 1-D click queries (ablation) | K-Prism interactive mode · interactive, ACDC+BUS | Dice(1): 85.44 -> 86.55 | Improved |
| initialize with Mode-1 semantic mask then refine with clicks (ablation) | K-Prism interactive · ACDC | NoC90: 2.67 -> 1.67 | Improved |
| initialize with Mode-2 in-context mask then refine with clicks (ablation) | K-Prism interactive · BUS | NoC90: 1.67 -> 1.17 | Improved |
| use in-context mode instead of semantic mode | K-Prism · KiTS | Dice: 78.70 -> 79.12 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- UniverSeg — UniverSeg: Universal Medical Image Segmentation (ICCV 2023) · better on 3 of 3 shared results
- MultiverSeg — MultiverSeg: Scalable Interactive Segmentation of Biomedical Imaging Datasets with In-Context Guidance (ICCV 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation (CVPR 2024 · similarity 0.90)
- Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.88)
- Show and Segment: Universal Medical Image Segmentation via In-Context Learning (CVPR 2025 · similarity 0.88)
- S2S2: Semantic Stacking for Robust Semantic Segmentation in Medical Imaging (AAAI 2025 · similarity 0.88)
- UniPro: Unified Multi-Mode Medical Image Segmentation from 2D Images to 3D Volumes via Propagation (NeurIPS 2026 · similarity 0.88)
- Multi-view Evidential Learning-based Medical Image Segmentation (AAAI 2025 · 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.
