Research map / Detection and segmentation / CNN features
Occlusion-Aware Instance Segmentation Via BiLayer Network Architectures
Lei Ke, Yu‐Wing Tai, Chi–Keung Tang
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 joint occluder contour and mask modeling in the first GCN (ablation) | BCNet (GCN) · COCO-OCC | mask AP: 29.04 -> 30.37 | Improved |
| add joint occluder contour and mask modeling in the first GCN (ablation) | BCNet (GCN) · COCO | mask AP: 32.65 -> 33.43 | Improved |
| add occluder contour on top of occluder mask (ablation) | BCNet (GCN) · COCO-OCC | mask AP: 29.65 -> 30.37 | Improved |
| add occlusion-aware guidance from first GCN to second GCN (ablation) | BCNet (GCN) · COCO-OCC | mask AP: 29.45 -> 30.37 | Improved |
| add occludee contour prediction in second GCN (ablation) | BCNet (GCN) · COCO-OCC | mask AP: 30.37 -> 30.68 | Improved |
| use bilayer instead of single GCN (ablation) | BCNet · COCO-OCC | mask AP: 29.63 -> 30.68 | Improved |
| use bilayer instead of single FCN (ablation) | BCNet · COCO-OCC | mask AP: 28.43 -> 30.12 | Improved |
| use GCN instead of FCN in single layer (ablation) | BCNet · COCO-OCC | mask AP: 28.43 -> 29.63 | Improved |
| use GCN instead of FCN in bilayer (ablation) | BCNet · COCO-OCC | mask AP: 30.12 -> 30.68 | Improved |
| add BCNet mask head to Faster R-CNN | Faster R-CNN · COCO-OCC | mask AP: 29.67 -> 31.71 | Improved |
| add bilayer transformer decoder instead of shared decoder with 200 queries (ablation) | Mask2Former · COCO-OCC | mask AP: 39.01 -> 40.17 | Improved |
| add occlusion guidance from first to second transformer decoder (ablation) | Transformer-based BCNet · COCO-OCC | mask AP: 40.17 -> 41.23 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- BlendMask — BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation (CVPR 2020) · better on 1 of 1 shared result
- CenterMask — CenterMask: Real-Time Anchor-Free Instance Segmentation (CVPR 2020) · better on 1 of 1 shared result
- PANet — PANet: Few-Shot Image Semantic Segmentation With Prototype Alignment (ICCV 2019) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Deep Occlusion-Aware Instance Segmentation With Overlapping BiLayers (CVPR 2021 · similarity 0.98)
- Robust Instance Segmentation Through Reasoning About Multi-Object Occlusion (CVPR 2021 · similarity 0.91)
- Occluded Video Instance Segmentation: A Benchmark (IJCV 2022 · similarity 0.90)
- BANet: Bidirectional Aggregation Network With Occlusion Handling for Panoptic Segmentation (CVPR 2020 · similarity 0.89)
- Boundary-Aware Instance Segmentation (CVPR 2017 · similarity 0.88)
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · 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.
