Research map / Detection and segmentation / Open-set and few-shot detection
ClevrTex: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentation
Laurynas Karazija, Iro Laina, Christian Rupprecht
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 |
|---|---|---|---|
| Train and evaluate on textured ClevrTex instead of CLEVR | SPAIR* · CLEVR vs ClevrTex | mIoU (%): 65.95 ± 4.02 -> 0.0 ± 0.0 | Hurt |
| Train and evaluate on textured ClevrTex instead of CLEVR | GNM · CLEVR vs ClevrTex | mIoU (%): 59.92 ± 3.72 -> 42.25 ± 0.18 | Hurt |
| Use varied textured backgrounds with simple colored objects instead of textured objects and background | SA · VarBG vs ClevrTex | mIoU (%): 22.58 ± 2.07 -> 62.57 | Improved |
| Use a consistent grass background instead of varied textured backgrounds | GNM · GrassBG vs ClevrTex | mIoU (%): 42.25 ± 0.18 -> 53.15 | Improved |
| Use a consistent grass background instead of varied textured backgrounds | SPACE · GrassBG vs ClevrTex | mIoU (%): 9.14 ± 3.46 -> 33.85 | Improved |
| Use textured objects on a plain background instead of textured objects and background | SPAIR* · PlainBG vs ClevrTex | mIoU (%): 0.0 ± 0.0 -> 39.32 | Mixed |
| Use varied textured backgrounds with simple colored objects instead of textured objects and background | SPAIR* · VarBG vs ClevrTex | mIoU (%): 0.0 ± 0.0 -> 0.00 | No clear effect |
| Use one consistent grass background instead of varied backgrounds | pixel-space methods · GrassBG vs ClevrTex | mIoU (%) | No clear effect |
| Use one common material on every object and the background | models trained on ClevrTex · CAMO | mIoU (%) | Hurt |
| Test on unseen materials and shapes | models trained on ClevrTex · OOD | mIoU (%) | Mixed |
Closest papers by meaning
- unMORE: Unsupervised Multi-Object Segmentation via Center-Boundary Reasoning (ICML 2025 · similarity 0.88)
- M^3-VOS: Multi-Phase, Multi-Transition, and Multi-Scenery Video Object Segmentation (CVPR 2025 · similarity 0.87)
- v-CLR: View-Consistent Learning for Open-World Instance Segmentation (CVPR 2025 · similarity 0.86)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.86)
- AGILE3D: Attention Guided Interactive Multi-object 3D Segmentation (ICLR 2024 · similarity 0.85)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.85)
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.
