Research map / Detection and segmentation / CNN features
IDRNet: Intervention-Driven Relation Network for Semantic Segmentation
Read the paperCheck 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 |
|---|---|---|---|
| add IDRNet to FCN | FCN · ADE20K val | mIoU: 36.96 -> 43.61 | Helped |
| add IDRNet to UPerNet | UPerNet · ADE20K val | mIoU: 43.02 -> 44.84 | Helped |
| add IDRNet to Faster R-CNN | Faster R-CNN · COCO detection | AP: 37.4 -> 38.2 | Helped |
| add dataset-level feature enhancement (ablation) | IDRNet · ADE20K val | mIoU: 42.82 -> 43.61 | Helped |
| add orthogonal-matrix feature enhancement (ablation) | IDRNet · ADE20K val | mIoU: 42.82 -> 42.99 | Helped |
| add mean relation matrix (ablation) | IDRNet · ADE20K val | mIoU: 42.74 -> 43.61 | Helped |
| add variance relation matrix (ablation) | IDRNet · ADE20K val | mIoU: 43.13 -> 43.61 | Helped |
| use balanced deletion instead of random deletion (ablation) | IDRNet · ADE20K val | mIoU: 42.90 -> 43.61 | Helped |
| use balanced deletion instead of random deletion (ablation) | IDRNet · PASCAL-Context val | mIoU: 52.92 -> 52.97 | No clear effect |
| update relation matrix by deletion diagnostics instead of back-propagation (ablation) | IDRNet · ADE20K val | mIoU: +3.26 mIoU | Helped |
| add IDRNet to DeeplabV3, train on Cityscapes, test on Dark Zurich | DeeplabV3 · Dark Zurich val | mIoU: 10.03 -> 13.66 | Helped |
| add IDRNet to UPerNet cross-domain | UPerNet · Dark Zurich val | mIoU: 11.06 -> 11.42 | No clear effect |
Closest papers by meaning
- Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2025 · similarity 0.85)
- SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation (CVPR 2025 · similarity 0.85)
- Integrating Low-Level Visual Cues for Enhanced Unsupervised Semantic Segmentation (AAAI 2025 · similarity 0.85)
- RelationVGGT : Visual Geometry Transformers for 3D Spatial Relation Segmentation (NeurIPS 2026 · similarity 0.85)
- Label-Efficient Few-Shot Semantic Segmentation with Unsupervised Meta-Training (AAAI 2024 · similarity 0.85)
- DeiSAM: Segment Anything with Deictic Prompting (NeurIPS 2024 · 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.