Research map / Detection and segmentation / CNN features
An End-To-End Network for Panoptic 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 the spatial ranking module (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.2 -> 39.0 | Helped |
| add the spatial ranking module (ablation) | OANet with ResNet-101 · COCO Panoptic validation | PQ: 38.8 -> 40.7 | Helped |
| replace instance training labels with panoptic non-overlapping instance annotations (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.2 -> 36.1 | Hurt |
| supervise the stuff branch on object classes as well as stuff classes (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 36.7 -> 37.2 | Helped |
| share backbone features between the stuff and instance branches (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 36.5 -> 37.0 | Helped |
| also share FPN skip-connection features (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.0 -> 37.2 | Helped |
| share backbone and FPN skip-connection features instead of no sharing (ablation) | OANet with ResNet-101 · COCO Panoptic validation | PQ: 38.2 -> 38.8 | Helped |
| set loss-balance λ to 0.25 rather than 0.2 (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 36.9 -> 37.2 | Helped |
| increase loss-balance λ from 0.25 to 0.33 (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.2 -> 36.9 | Hurt |
| increase loss-balance λ from 0.25 to 0.50 (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.2 -> 36.5 | Hurt |
| increase loss-balance λ from 0.25 to 0.75 (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ: 37.2 -> 35.3 | Hurt |
| set loss-balance λ to 1.0 (ablation) | OANet with ResNet-50 · COCO Panoptic validation | PQ | Hurt |
Closest papers by meaning
- BANet: Bidirectional Aggregation Network With Occlusion Handling for Panoptic Segmentation (CVPR 2020 · similarity 0.92)
- Learning Instance Occlusion for Panoptic Segmentation (CVPR 2020 · similarity 0.90)
- SOGNet: Scene Overlap Graph Network for Panoptic Segmentation (AAAI 2020 · similarity 0.89)
- Revisiting Open-Set Panoptic Segmentation (AAAI 2024 · similarity 0.87)
- PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic Segmentation (CVPR 2024 · similarity 0.87)
- Scene-Centric Unsupervised Panoptic Segmentation (CVPR 2025 · similarity 0.86)
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.