Research map / Detection and segmentation / CNN features
Context Prior for Scene Segmentation
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 |
|---|---|---|---|
| Add Context Prior Layer to ResNet-50 | ResNet-50 scene segmentation model · ADE20K validation | mIoU: 36.24 -> 43.92 | Helped |
| Add auxiliary loss to the dilated ResNet-50 baseline (ablation) | ResNet-50 scene segmentation model · ADE20K validation | mIoU: 34.38%→36.24% | Helped |
| Add IntraPrior branch trained with BCE (ablation) | ResNet-50 baseline · ADE20K validation | mIoU: 42.34% | Helped |
| Add InterPrior branch trained with BCE (ablation) | ResNet-50 baseline · ADE20K validation | mIoU: 42.88% | Helped |
| Use Affinity Loss rather than BCE for IntraPrior (ablation) | IntraPrior branch · ADE20K validation | mIoU: +0.4% | Helped |
| Use Affinity Loss rather than BCE for InterPrior (ablation) | InterPrior branch · ADE20K validation | mIoU: +0.55% | Helped |
| Add multi-scale and flip testing (ablation) | ResNet-50 CPNet · ADE20K validation | mIoU: 43.92% -> 44.46% | Helped |
| Use ResNet-101 instead of ResNet-50 (ablation) | CPNet · ADE20K validation | mIoU: 44.46 -> 45.39 | Helped |
| Add Context Prior to the Aggregation Module (ablation) | ResNet-50 model with Aggregation Module, kernel size 11 · ADE20K validation | mIoU: 41.51 -> 43.92 | Helped |
| Increase the kernel size to 11 (ablation) | Context Prior Aggregation Module · ADE20K validation | mIoU: 0.2; 0.95; 1.51; 0.82; 2.41; 0.2; 0.36 | Mixed |
| Replace Aggregation Module with PPM and retain Context Prior (ablation) | Context Prior Layer · ADE20K validation | mIoU: ↑1.06 | Helped |
| Replace Aggregation Module with ASPP and retain Context Prior (ablation) | Context Prior Layer · ADE20K validation | mIoU: ↑2.3 | Helped |
Closest papers by meaning
- Context-Reinforced Semantic Segmentation (CVPR 2019 · similarity 0.89)
- Context Encoding for Semantic Segmentation (CVPR 2018 · similarity 0.89)
- Contextrast: Contextual Contrastive Learning for Semantic Segmentation (CVPR 2024 · similarity 0.88)
- Adaptive Pyramid Context Network for Semantic Segmentation (CVPR 2019 · similarity 0.88)
- Explore In-Context Segmentation via Latent Diffusion Models (AAAI 2025 · similarity 0.87)
- Context Encoders: Feature Learning by Inpainting (CVPR 2016 · 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.