Research map / Detection and segmentation / CNN features
Deep Occlusion-Aware Instance Segmentation With Overlapping BiLayers
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 explicit occluder mask and contour modeling (ablation) | BCNet · COCO-OCC | AP: 29.04 -> 30.37 | Helped |
| Add occluder mask and contour modeling (ablation) | BCNet · COCO | AP: 32.65 -> 33.43 | Helped |
| Add occluder mask modeling without contours (ablation) | BCNet · COCO-OCC | AP: 29.04 -> 29.65 | Helped |
| Add occluder contour modeling without masks (ablation) | BCNet · COCO-OCC | AP: 29.04 -> 30.18 | Helped |
| Add occlusion-aware guidance from the first layer (ablation) | BCNet occludee head · COCO-OCC | AP: 29.45 -> 30.37 | Helped |
| Add occludee contour prediction (ablation) | Guided BCNet occludee head · COCO-OCC | AP: 30.37 -> 30.68 | Helped |
| Replace single-layer GCN with bilayer GCN (ablation) | BCNet · COCO-OCC | AP: 29.63 -> 30.68 | Helped |
| Replace single-layer FCN with bilayer FCN (ablation) | BCNet · COCO-OCC | AP: 28.43 -> 30.12 | Helped |
| Use GCN rather than FCN (ablation) | Single-layer and bilayer mask heads · COCO-OCC | AP: 28.43 -> 29.63; 30.12 -> 30.68 | Helped |
| Train on synthetic occlusion data (ablation) | BCNet · COCO-OCC | AP: 31.71 -> 32.89 | Helped |
| Add BCNet's bilayer mask head to the detector baseline | FCOS · COCO-OCC | AP: 28.43 -> 30.68 | Helped |
| Add BCNet's bilayer mask head to the detector baseline | Faster R-CNN · COCO-OCC | AP: 29.67 -> 31.71 | Helped |
Closest papers by meaning
- Robust Instance Segmentation Through Reasoning About Multi-Object Occlusion (CVPR 2021 · similarity 0.91)
- Learning Instance Occlusion for Panoptic Segmentation (CVPR 2020 · similarity 0.88)
- BANet: Bidirectional Aggregation Network With Occlusion Handling for Panoptic Segmentation (CVPR 2020 · similarity 0.88)
- An End-To-End Network for Panoptic Segmentation (CVPR 2019 · similarity 0.87)
- In the Shadows, Shape Priors Shine: Using Occlusion to Improve Multi-Region Segmentation (CVPR 2016 · similarity 0.87)
- Learning to Generate Multiple Objects from Dense and Occluded Layouts (NeurIPS 2026 · 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.