Research map / Detection and segmentation / CNN features
BANet: Bidirectional Aggregation Network With Occlusion Handling 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 |
|---|---|---|---|
| Use BANet instead of AUNet | COCO val, ResNet-50-FPN panoptic segmentation · COCO val | PQ: 39.6 -> 41.1 | Helped |
| Use BANet with deformable convolution instead of UpsNet | COCO val, ResNet-50-FPN panoptic segmentation · COCO val | PQ: 42.5 -> 43.0 | Helped |
| Use BANet with deformable convolution instead of UpsNet | COCO test-dev, DCN-101-FPN panoptic segmentation · COCO test-dev | PQ: 46.6 -> 47.3 | Helped |
| Add SIM to the baseline (ablation) | BANet · COCO val | PQ: 39.1 -> 39.5 | Helped |
| Add OCM to SIM (ablation) | BANet · COCO val | PQ: 39.5 -> 40.0 | Helped |
| Add S2I to SIM and OCM (ablation) | BANet · COCO val | RQTh: +0.4% | Helped |
| Add deformable convolution to the semantic head (ablation) | BANet with SIM, OCM, and S2I · COCO val | PQ: 40.3 -> 41.8 | Helped |
| Add occlusion handling instead of heuristic fusion (ablation) | BANet with deformable convolution · COCO val | PQTh: 48.5 -> 50.5 | Helped |
| Use ResNet-101-FPN instead of ResNet-50-FPN (ablation) | BANet with occlusion handling · COCO val | PQ: 43.0 -> 44.0 | Helped |
| Replace RoIInlay with RoIUpsample (ablation) | Model with SIM · COCO val | SQ: 78.0 -> 77.0 | Hurt |
| Use ground-truth overlap relationships for occlusion resolution (ablation) | BANet · COCO val | PQTh: 50.5 -> 53.2 | Helped |
Closest papers by meaning
- An End-To-End Network for Panoptic Segmentation (CVPR 2019 · similarity 0.92)
- Learning Instance Occlusion for Panoptic Segmentation (CVPR 2020 · similarity 0.90)
- Bidirectional Graph Reasoning Network for Panoptic Segmentation (CVPR 2020 · similarity 0.90)
- Auto-Panoptic: Cooperative Multi-Component Architecture Search for Panoptic Segmentation (NeurIPS 2020 · similarity 0.89)
- Combinatorial Optimization for Panoptic Segmentation: A Fully Differentiable Approach (NeurIPS 2021 · similarity 0.88)
- Panoptic Feature Pyramid Networks (CVPR 2019 · similarity 0.88)
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.