Research map / Detection and segmentation / Weakly supervised detection and segmentation
SAMIX: Reinforcing SAM2 with Semantic Adapter and Reference Selecting Policy for Mix-Supervised Segmentation
Qiang Hu, Jiajie Wei, Zhenyu Yi, Zhifen Yan, Yingjie Guo, Hongkuan Shi, Ge-Peng Ji, Qiang Li, Zhiwei Wang
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 semantic adapter to SAM2 image encoder (without SPNet) (ablation) | SAMIX baseline · PASCAL VOC 2012 val | mIoU: 66.5 -> 72.5 | Improved |
| add SPNet reference selection (without adapter) (ablation) | SAMIX baseline · PASCAL VOC 2012 val | mIoU: 66.5 -> 68.7 | Improved |
| add both adapter and SPNet (ablation) | SAMIX baseline · PASCAL VOC 2012 val | mIoU: 66.5 -> 75.9 | Improved |
| AdaptFormer as adapter instead of naive adapter/LoRA/Med-SA (ablation) | SAMIX semantic adapter · PASCAL VOC 2012 val | mIoU: 75.5 (LoRA) -> 75.9 | Improved |
| use SPNet instead of random reference selection (ablation) | SAMIX · PASCAL VOC 2012 val | mIoU: 70.4 -> 75.9 | Improved |
| use SPNet instead of heuristic sim(F,E) selection (ablation) | SAMIX · PASCAL VOC 2012 val | mIoU: 72.7 -> 75.9 | Improved |
| add Hierarchical Guidance Principle to SPNet (ablation) | SPNet · PASCAL VOC 2012 val | mIoU: 74.4 -> 75.9 | Improved |
| add momentum-updated memory (ablation) | SPNet · PASCAL VOC 2012 val | mIoU: 75.3 -> 75.9 | Improved |
| add consistency reward to quality reward (ablation) | SPNet RL rewards · PASCAL VOC 2012 val | mIoU: 73.5 -> 75.0 | Improved |
| add supervision reward on top of quality and consistency (ablation) | SPNet RL rewards · PASCAL VOC 2012 val | mIoU: 75.0 -> 75.9 | Improved |
| replace SAMIX adapter and SPNet by baseline on COD10K (ablation) | SAMIX baseline · COD10K COD | Sm: 71.8 -> 79.1 | Improved |
| switch SAMIX from weakly-supervised to mix-supervision under equal annotation cost | SAMIX · PASCAL VOC 2012 (equal annotation cost) | mIoU: 71.5 -> 73.7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- WISH — WISH: Weakly Supervised Instance Segmentation using Heterogeneous Labels (CVPR 2025) · better on 5 of 5 shared results
- SAM-COD — SAM-COD: SAM-guided Unified Framework for Weakly-Supervised Camouflaged Object Detection (ECCV 2024) · better on 5 of 5 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
- Semantic-aware SAM for Point-Prompted Instance Segmentation (CVPR 2024 · similarity 0.88)
- SAM-CP: Marrying SAM with Composable Prompts for Versatile Segmentation (ICLR 2025 · similarity 0.87)
- From SAM to CAMs: Exploring Segment Anything Model for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.87)
- SANSA: Unleashing the Hidden Semantics in SAM2 for Few-Shot Segmentation (NeurIPS 2025 · similarity 0.87)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 2024 · similarity 0.86)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · 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.
