Research map / Detection and segmentation / Open-set and few-shot detection
In the Shadows, Shape Priors Shine: Using Occlusion to Improve Multi-Region 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 |
|---|---|---|---|
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 2 objects, region2 | AP: 89.35 -> 94.04 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 2 objects, region2 | IoU: 80.95 -> 89.52 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 3 objects, region2 | AP: 92.88 -> 96.25 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 3 objects, region2 | IoU: 87.24 -> 93.24 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 3 objects, region3 | AP: 82.87 -> 86.98 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic motorbikes, 3 objects, region3 | IoU: 71.39 -> 78.41 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic horses, 2 objects, region2 | AP: 82.58 -> 87.62 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic horses, 2 objects, region2 | IoU: 69.12 -> 77.96 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic horses, 3 objects, region3 | AP: 69.97 -> 71.39 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic horses, 3 objects, region3 | IoU: 47.91 -> 50.36 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic Caltech-101 multiclass, 3 objects, region2 | AP: 85.59 -> 92.04 | Helped |
| Jointly segment overlapping regions while applying the shape prior only to visible-pixel data agreement, instead of segmenting each object separately | Shape-prior segmentation · Synthetic Caltech-101 multiclass, 3 objects, region3 | IoU: 57.96 -> 66.35 | Helped |
Closest papers by meaning
- Robust Instance Segmentation Through Reasoning About Multi-Object Occlusion (CVPR 2021 · similarity 0.90)
- Using Diffusion Priors for Video Amodal Segmentation (CVPR 2025 · similarity 0.86)
- unMORE: Unsupervised Multi-Object Segmentation via Center-Boundary Reasoning (ICML 2025 · similarity 0.85)
- Large Occluded Human Image Completion via Image-Prior Cooperating (AAAI 2024 · similarity 0.85)
- Robust 3D Shape Reconstruction in Zero-Shot from a Single Image in the Wild (CVPR 2025 · similarity 0.85)
- Multi-Part Object Representations via Graph Structures and Co-Part Discovery (NeurIPS 2026 · 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.