Research map / Graph learning / Graph self-supervised learning
SEMIR: Semantic Minor-Induced Representation Learning on Graphs for Visual Segmentation
Luke James Miller, Yugyung Lee
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 edge contraction (ablation) | SEMIR · BraTS ET | Dice: 0.441 -> 0.894 | Improved |
| add edge deletion (ablation) | SEMIR · BraTS ET | Dice: 0.719 -> 0.894 | Improved |
| add node deletion (ablation) | SEMIR · BraTS ET | Dice: 0.812 -> 0.894 | Improved |
| learn minor parameters with few-shot optimization instead of manual tuning (ablation) | SEMIR · BraTS ET | Dice: 0.837 -> 0.894 | Improved |
| learn minor parameters with few-shot optimization instead of manual tuning (ablation) | SEMIR · NWPU VHR-10 | small-object IoU: 0.763 -> 0.789 | Improved |
| add relational edge features (ablation) | SEMIR · BraTS ET | Dice: 0.725 -> 0.894 | Improved |
| add spatial features (compactness, elongation) (ablation) | SEMIR · BraTS ET | Dice: 0.661 -> 0.894 | Improved |
| increase connectivity from 6 to 18 (ablation) | SEMIR · NWPU VHR-10 | small-object IoU: +3-5% IoU at 2.5x runtime | Mixed |
| SEMIR binary graph-minor model versus nnU-Net | SEMIR · BraTS ET/TC, KiTS tumor, LiTS tumor | Dice: KiTS T 0.720 -> 0.819; LiTS T 0.733 -> 0.891 | Improved |
| SEMIR versus best published methods on aggregate regions | SEMIR · BraTS WT, LiTS LT | Dice: BraTS WT 0.943 (GTMamba) vs 0.920; LiTS LT 0.977 (Lgma-net) vs 0.959 | Hurt |
Closest papers by meaning
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.86)
- Beyond Pixel-wise Supervision: Local Structure Regularization for Semantic Segmentation (NeurIPS 2026 · similarity 0.86)
- Graph-BAS3Net: Boundary-Aware Semi-Supervised Segmentation Network With Bilateral Graph Convolution (ICCV 2021 · similarity 0.86)
- SAM-Aware Graph Prompt Reasoning Network for Cross-Domain Few-Shot Segmentation (AAAI 2025 · similarity 0.86)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.86)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.85)
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.
