Research map / Detection and segmentation / Open-set and few-shot detection
CObL: Toward Zero-Shot Ordinal Layering without User Prompting
Aneel Damaraju, Dean Hazineh, Todd Zickler
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 MiDaS depth cue to the conditioning adapter (ablation) | CObL · TableTop ablation | LPIPS: LPIPS increases by 4% without it | Improved |
| keep Stable Diffusion UNet frozen instead of unfreezing it (ablation) | CObL · TableTop | LPIPS: 9% increase in LPIPS when unfrozen | Improved |
| add prior score matching and compositional guidance at inference (ablation) | CObL · TableTop | LPIPS: 8% LPIPS increase without guidance | Improved |
| increase number of objects in scene beyond four (ablation) | CObL · TableTop | LPIPS and CLIP | Hurt |
| add non-differentiable permute, erase and sort updates during sampling (ablation) | CObL · TableTop | evaluation metrics | No clear effect |
| use CObL instead of pix2gestalt with oracle masks | amodal completion · TableTop (60 scenes, 2-4 objects) | LPIPS best: .128 -> .094 | Improved |
| use CObL instead of pix2gestalt with oracle masks | amodal completion · TableTop (60 scenes, 2-4 objects) | CLIP best: .889 -> .935 | Improved |
| use CObL instead of SDXL Inpainting with oracle masks | amodal completion · TableTop (60 scenes, 2-4 objects) | LPIPS best: .373 -> .094 | Improved |
| use CObL instead of LaMa with oracle masks | amodal completion · TableTop (60 scenes, 2-4 objects) | LPIPS best: .113 -> .094 | Improved |
| use CObL instead of pix2gestalt | amodal completion · TableTop (60 scenes, 2-4 objects) | LPIPS average: .153 -> .122 | Improved |
| use CObL instead of Mask2Former for visible segmentation | panoptic segmentation · TableTop visible segmentation | ARI: 66.3% -> 83.5% | Improved |
Closest papers by meaning
- ObjectGS: Object-aware Scene Reconstruction and Scene Understanding via Gaussian Splatting (ICCV 2025 · similarity 0.88)
- DORSal: Diffusion for Object-centric Representations of Scenes $\textit{et al.}$ (ICLR 2024 · similarity 0.87)
- MOVIS: Enhancing Multi-Object Novel View Synthesis for Indoor Scenes (CVPR 2025 · similarity 0.86)
- Object-X: Learning to Reconstruct Multi-Modal 3D Object Representations (NeurIPS 2025 · similarity 0.86)
- Learning to Generate Multiple Objects from Dense and Occluded Layouts (NeurIPS 2026 · similarity 0.86)
- GLASS: Guided Latent Slot Diffusion for Object-Centric Learning (CVPR 2025 · 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.
