Research map / Detection and segmentation / Open-set and few-shot detection
Towards Fine-grained Interactive Segmentation in Images and Videos
Yuan Yao, Qiushi Yang, Miaomiao Cui, Liefeng Bo
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 Localization Augment (LA) to SAM2 (ablation) | SAM2 · Average of 4 fine-grained datasets | mIoU: 78.5 -> 85.0 | Improved |
| add Localization Augment (LA) to SAM2 (ablation) | SAM2 · Average of 4 fine-grained datasets | mBIoU: 71.5 -> 67.4 | Hurt |
| add Prompt Retargeting (PR) to LA+MR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mIoU: 92.7 -> 93.1 | Improved |
| add Prompt Retargeting (PR) to LA+MR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mBIoU: 88.3 -> 88.9 | Improved |
| add Mask Refinement (MR) to LA+PR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mIoU: 91.4 -> 93.1 | Improved |
| add Mask Refinement (MR) to LA+PR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mBIoU: 86.2 -> 88.9 | Improved |
| add Localization Augment (LA) to PR+MR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mIoU: 91.1 -> 93.1 | Improved |
| add Localization Augment (LA) to PR+MR (ablation) | SAM2Refiner · Average of 4 fine-grained datasets | mBIoU: 86.0 -> 88.9 | Improved |
| use SAM2Refiner (hiera_l) instead of HQ-SAM (vit_h) | SAM2Refiner · Average of 4 fine-grained datasets | mIoU: 88.2 -> 93.1 | Improved |
| use SAM2Refiner instead of SAM2 on ThinObject | SAM2Refiner · ThinObject | mIoU: 77.8 -> 95.9 | Improved |
| use SAM2Refiner instead of SAM2 on LVOS v1 | SAM2Refiner · LVOS v1 | J&F: 77.2 -> 80.9 | Improved |
| use SAM2Refiner instead of HQ-SAM2 on LVOS v2 | SAM2Refiner · LVOS v2 | J&F: 79.1 -> 79.9 | Improved |
Closest papers by meaning
- SAM-REF: Introducing Image-Prompt Synergy during Interaction for Detail Enhancement in the Segment Anything Model (CVPR 2025 · similarity 0.90)
- SAM 2: Segment Anything in Images and Videos (ICLR 2025 · similarity 0.90)
- SAM-I2V: Upgrading SAM to Support Promptable Video Segmentation with Less than 0.2% Training Cost (CVPR 2025 · similarity 0.89)
- FocSAM: Delving Deeply into Focused Objects in Segmenting Anything (CVPR 2024 · similarity 0.89)
- Rethinking Interactive Image Segmentation with Low Latency High Quality and Diverse Prompts (CVPR 2024 · similarity 0.89)
- AoP-SAM: Automation of Prompts for Efficient Segmentation (AAAI 2025 · similarity 0.89)
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.
