Research map / Detection and segmentation / CNN features
Learning Affinity via Spatial Propagation Networks
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 |
|---|---|---|---|
| Use three-way rather than one-way propagation connections (ablation) | SPN on HELEN face parsing · HELEN | overall IoU: 87.09 -> 89.30 | Helped |
| Add three-way SPN | FCN-8s · PASCAL VOC 2012 val | mean IoU: 65.51 -> 69.86 | Helped |
| Use one-way propagation instead of three-way propagation (ablation) | FCN-8s + SPN · PASCAL VOC 2012 val | mean IoU: 69.86 -> 60.95 | Hurt |
| Add three-way SPN | Deeplab VGG · PASCAL VOC 2012 val | mean IoU: 68.97 -> 73.12 | Helped |
| Use one-way propagation instead of three-way propagation (ablation) | Deeplab VGG + SPN · PASCAL VOC 2012 val | mean IoU: 73.12 -> 64.42 | Hurt |
| Add three-way SPN | Deeplab ResNet-101 · PASCAL VOC 2012 val | mean IoU: 76.46 -> 79.76 | Helped |
| Use one-way propagation instead of three-way propagation (ablation) | Deeplab ResNet-101 + SPN · PASCAL VOC 2012 val | mean IoU: 79.76 -> 72.02 | Hurt |
| Replace dense CRF refinement with SPN | Deeplab VGG · PASCAL VOC 2012 val | mean IoU: 71.57 -> 73.12 | Helped |
| Replace dense CRF refinement with SPN | Deeplab ResNet · PASCAL VOC 2012 val | mean IoU: 77.69 -> 79.76 | Helped |
| Add three-way SPN | Front end dilated-convolution model · PASCAL VOC 2012 val | mean IoU: 69.75 -> 73.14 | Helped |
| Add three-way SPN on top of Context | Front end + Context · PASCAL VOC 2012 val | mean IoU: 71.86 -> 75.28 | Helped |
Closest papers by meaning
- Learning Pixel-Level Semantic Affinity With Image-Level Supervision for Weakly Supervised Semantic Segmentation (CVPR 2018 · similarity 0.87)
- Learning Affinity with Hyperbolic Representation for Spatial Propagation (ICML 2023 · similarity 0.87)
- Affinity Graph Supervision for Visual Recognition (CVPR 2020 · similarity 0.87)
- Label-efficient Segmentation via Affinity Propagation (NeurIPS 2023 · similarity 0.86)
- Deciphering 'What' and 'Where' Visual Pathways from Spectral Clustering of Layer-Distributed Neural Representations (CVPR 2024 · similarity 0.84)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.84)
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.