Research map / Detection and segmentation / Open-set and few-shot detection
Ensemble Foreground Management for Unsupervised Object Discovery
Ziling Wu, Armaghan Moemeni, Praminda Caleb-Solly
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 aggregated unit voters as foreground prior (ablation) | TokenCut · COCO20K | CorLoc: 58.8 -> 62.0 | Improved |
| add corner prior to aggregated UVs (ablation) | TokenCut+aU · COCO20K | CorLoc: 62.0 -> 62.1 | Improved |
| add distillation (UnionSeg) on top of UnionCut (ablation) | TokenCut · VOC07 | CorLoc: 69.2 -> 69.7 | Improved |
| add distillation (UnionSeg) on top of UnionCut (ablation) | TokenCut · COCO20K | CorLoc: 62.1 -> 62.6 | Improved |
| add UnionSeg foreground prior | LOST · COCO20K | CorLoc: 50.7 -> 54.9 | Improved |
| add UnionCut foreground prior | TokenCut · COCO20K | CorLoc: 58.8 -> 62.1 | Improved |
| add UnionCut or UnionSeg to the strong FOUND baseline | FOUND · VOC07 | CorLoc: 72.5 -> 72.6 (UnionCut), 72.3 (UnionSeg) | Mixed |
| add UnionSeg foreground prior | CutLER · VOC12 | CorLoc: 69.5 -> 71.2 | Improved |
| add UnionSeg foreground prior | CutLER · DUTS-TE saliency | maxF_beta: 54.2 -> 72.3 | Improved |
| add UnionCut foreground prior | TokenCut · DUTS-TE saliency | maxF_beta: 75.5 -> 80.2 | Improved |
| use UnionSeg pseudo-labels for CutLER training | CutLER · COCO20K instance segmentation | AP mask: 9.2 -> 10.0 | Improved |
| distill UnionCut into UnionSeg for efficiency | UnionCut · foreground union speed | FPS: 0.1 -> 125 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TokenCut — TokenCut: Segmenting Objects in Images and Videos With Self-Supervised Transformer and Normalized Cut (TPAMI 2023) · better on 4 of 4 shared results
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
- Large-Scale Unsupervised Object Discovery (NeurIPS 2021 · similarity 0.89)
- UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classes (NeurIPS 2024 · similarity 0.87)
- unMORE: Unsupervised Multi-Object Segmentation via Center-Boundary Reasoning (ICML 2025 · similarity 0.87)
- FOCUS: Towards Universal Foreground Segmentation (AAAI 2025 · similarity 0.86)
- UN-DETR: Promoting Objectness Learning via Joint Supervision for Unknown Object Detection (AAAI 2025 · similarity 0.85)
- CuVLER: Enhanced Unsupervised Object Discoveries through Exhaustive Self-Supervised Transformers (CVPR 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.
