Research map / 3D vision / 3D generation
FoveaNet: Perspective-Aware Urban Scene Parsing
Xin Li, Zequn Jie, Wei Wang, Changsong Liu, Jimei Yang, Xiaohui Shen, Zhe Lin, Qiang Chen, Shuicheng Yan, Jiashi Feng
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 PEN-localized fovea region instead of a fixed one (ablation) | FoveaNet parsing net · Cityscapes val | iIoU class: 54.6 -> 57.4 (margin of 2.8%) | Improved |
| add PEN fovea region with scale normalization (ablation) | FCN baseline · Cityscapes val | iIoU class: 51.8 -> 57.4 (5.6%) | Improved |
| add PEN fovea region with scale normalization (ablation) | FCN baseline · Cityscapes val | iIoU category: 72.8 -> 76.6 (3.8%) | Improved |
| add a fixed fovea region from the global average (ablation) | FCN baseline · Cityscapes val | iIoU class: 51.8 -> 54.6 | Improved |
| add normal dense CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class: 73.9 -> 73.5 | Hurt |
| add normal dense CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class of truck, bus, train: up to 10.4% | Improved |
| add perspective-aware CRFs instead of normal CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class: 73.5 -> 75.9 | Improved |
| add perspective-aware CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class: 73.9 -> 75.9 | Improved |
| add perspective-aware CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class of truck, bus, train: +15.7%, 6.0%, 12.3% | Improved |
| use perspective heatmap instead of predicted depth in the CRFs (ablation) | perspective-aware CRFs · Cityscapes val | IoU class: 74.3 -> 75.9 | Improved |
| add depth-aware CRFs instead of normal CRFs (ablation) | FCN + PEN fovea region · Cityscapes val | IoU class: 73.5 -> 74.3 | Improved |
| FoveaNet vs plain FCN | ResNet-101 FCN · Cityscapes test | IoU class: 71.3 -> 74.1 | Improved |
Closest papers by meaning
- Perspective-Adaptive Convolutions for Scene Parsing (TPAMI 2019 · similarity 0.89)
- Capturing Omni-Range Context for Omnidirectional Segmentation (CVPR 2021 · similarity 0.88)
- PSANet: Point-wise Spatial Attention Network for Scene Parsing (ECCV 2018 · similarity 0.88)
- The Cityscapes Dataset for Semantic Urban Scene Understanding (CVPR 2016 · similarity 0.87)
- Learning Content-Enhanced Mask Transformer for Domain Generalized Urban-Scene Segmentation (AAAI 2024 · similarity 0.85)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · 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.
