Research map / Detection and segmentation / Open-set and few-shot detection
FoundYou: A Unified Model for Personalized Segmentation and Retrieval
Gabriele Trivigno, Marcos Alfaro Perez, Claudia Cuttano, Gabriele Berton, Luis Payá, Carlo Masone
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 self-distillation on dense mask outputs (ablation) | FoundYou adapter · PerMIS | mIoU: 56.7 -> 62.6 | Improved |
| add self-distillation on dense mask outputs (ablation) | FoundYou adapter · ILIAS | mAP: 32.7 -> 32.5 | No clear effect |
| use lightweight adapter instead of full fine-tuning (ablation) | SAM 2 · PerMIS | mIoU: 27.8 -> 62.6 | Improved |
| use lightweight adapter instead of frozen SAM 2 (ablation) | SAM 2 · ILIAS | mAP: 3.5 -> 32.5 | Improved |
| use AdaptFormer instead of LoRA (ablation) | FoundYou adapter · ILIAS | mAP: 28.9 -> 32.5 | Improved |
| use 0.5x bottleneck instead of 0.3x (ablation) | AdaptFormer · ILIAS | mAP: 29.8 -> 32.5 | Improved |
| adapt late stages 2-3 instead of none (ablation) | Image Encoder · ILIAS | mAP: 20.4 -> 32.5 | Improved |
| adapt all stages 0-3 instead of late stages only (ablation) | Image Encoder · ILIAS | mAP: 32.5 -> 31.8 | Hurt |
| increase references from 1 to 4 | FoundYou · ILIAS few-shot | mAP: 30.1 -> 34.2 | Improved |
| increase re-ranking set K from 10 to 1000 | FoundYou · ILIAS | mAP: 23.4 -> 32.5 | Improved |
| increase re-ranking set K from 100 to 1000 | RoMa v2 · ILIAS | mAP: 27.3 -> 26.9 | Hurt |
| use point prompt instead of mask prompt | FoundYou · PerMIS | mIoU: 62.6 -> 58.6 | Hurt |
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.
- SegGPT — SegGPT: Towards Segmenting Everything in Context (ICCV 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
- Where's Waldo: Diffusion Features For Personalized Segmentation and Retrieval (NeurIPS 2024 · similarity 0.89)
- Personalize Segment Anything Model with One Shot (ICLR 2024 · similarity 0.87)
- SANSA: Unleashing the Hidden Semantics in SAM2 for Few-Shot Segmentation (NeurIPS 2025 · similarity 0.87)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.86)
- The Devil is in the Object Boundary: Towards Annotation-free Instance Segmentation using Foundation Models (ICLR 2024 · similarity 0.86)
- Perceive Anything: Recognize, Explain, Caption, and Segment Anything in Images and Videos (NeurIPS 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.
