Research map / Detection and segmentation / Open-set and few-shot detection
Benchmarking and Analysis of Unsupervised Object Segmentation from Real-World Single Images
Yafei Yang, Bo Yang
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 |
|---|---|---|---|
| erase object color gradients by replacing each object with its mean color (ablation) | AIR, MONet, IODINE, SlotAtt · real-world datasets (MOViC, YCB, ScanNet, COCO) | segmentation performance (ARI/AP) | Improved |
| make object shapes convex (reduce concavity) (ablation) | AIR, MONet, IODINE, SlotAtt · real-world datasets | segmentation performance | Improved |
| replace object textures with distinctive textures (lower inter-object color similarity) (ablation) | AIR, MONet, IODINE, SlotAtt · real-world datasets | segmentation performance | Improved |
| normalize object sizes within images (ablation) | AIR · real-world datasets | segmentation performance | Improved |
| normalize object sizes within images (ablation) | MONet, IODINE, SlotAtt · real-world datasets | segmentation performance | No clear effect |
| combine object color, shape and scene color ablation (C+S+T) (ablation) | SlotAtt · real-world datasets | segmentation performance | Improved |
| combine object color, shape and size-normalization ablation (C+S+U) (ablation) | AIR, MONet, IODINE, SlotAtt · real-world datasets | segmentation performance | Mixed |
| ablate all four object and scene factors (C+S+T+U) (ablation) | all four baselines · real-world datasets | segmentation performance | Improved |
| replace background with average color (remove background color gradient) (ablation) | IODINE, SlotAtt · real-world datasets with backgrounds | background recall | Improved |
| replace background texture with a distinctive one against foreground (ablation) | IODINE, SlotAtt · real-world datasets with backgrounds | background recall | Improved |
| make background contours convex (regular) (ablation) | IODINE, SlotAtt · real-world datasets with backgrounds | segmentation and background recall | No clear effect |
| use simple, discriminative backgrounds to help object segmentation (ablation) | IODINE, SlotAtt · real-world datasets with backgrounds | foreground object segmentation | No clear effect |
Closest papers by meaning
- Promising or Elusive? Unsupervised Object Segmentation from Real-world Single Images (NeurIPS 2022 · similarity 0.99)
- Unsupervised Object Segmentation in Video by Efficient Selection of Highly Probable Positive Features (ICCV 2017 · similarity 0.90)
- unMORE: Unsupervised Multi-Object Segmentation via Center-Boundary Reasoning (ICML 2025 · similarity 0.89)
- Unsupervised Object Localization: Observing the Background To Discover Objects (CVPR 2023 · similarity 0.88)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.87)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.87)
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.
