Research map / Detection and segmentation / Open-set and few-shot detection
BREPS: Bounding-Box Robustness Evaluation of Promptable Segmentation
Andrey Moskalenko, Danil Kuznetsov, Irina Dudko, Anastasiia Iasakova, Nikita Boldyrev, Denis Shepelev, Andrei Spiridonov, Andrey Kuznetsov, Vlad Shakhuro
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 |
|---|---|---|---|
| replace tight box with adversarial (min) box under realism constraint | SAM ViT-B · COCO-MVal | IoU: 86.95 -> 67.73 | Hurt |
| replace tight box with adversarial (min) box | SAM 2.1 Hiera-B+ · ADE20K | IoU: 79.01 -> 22.32 | Hurt |
| replace tight box with adversarial (min) box | SAM-HQ ViT-H · COCO-MVal | IoU: 87.67 -> 76.73 | Hurt |
| optimize box to maximize IoU instead of using tight box | SAM ViT-B · COCO-MVal | IoU: 86.95 -> 89.76 | Improved |
| use real-user boxes instead of tight boxes | SAM ViT-B · COCO | IoU: 86.95 -> 80.83 | Hurt |
| use mobile instead of desktop for drawing boxes | segmentation models · real-user study | IoU | Hurt |
| add bbox quality (CIoU-Loss) as predictor of segmentation IoU | SAM ViT-B · real-user boxes | correlation: SRCC 0.28 | No clear effect |
| scale up RobustSAM backbone from ViT-B to ViT-L | RobustSAM · COCO-MVal | IoU-Tight: 77.66 -> 52.06 | Hurt |
| scale SAM-HQ backbone ViT-B to ViT-H | SAM-HQ · COCO-MVal | IoU-Delta: 19.69 -> 12.79 | Improved |
Closest papers by meaning
- Stable Segment Anything Model (ICLR 2025 · similarity 0.87)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · similarity 0.86)
- AoP-SAM: Automation of Prompts for Efficient Segmentation (AAAI 2025 · similarity 0.86)
- Rethinking Interactive Image Segmentation with Low Latency High Quality and Diverse Prompts (CVPR 2024 · similarity 0.86)
- TETRIS: Towards Exploring the Robustness of Interactive Segmentation (AAAI 2024 · similarity 0.85)
- RobustSAM: Segment Anything Robustly on Degraded Images (CVPR 2024 · similarity 0.84)
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.
