Research map / Detection and segmentation / Weakly supervised detection and segmentation
Segment Concealed Objects With Incomplete Supervision
Chunming He, Kai Li, Yachao Zhang, Ziyun Yang, Youwei Pang, Longxiang Tang, Chengyu Fang, Yulun Zhang, Linghe Kong, Xiu Li, Sina Farsiu
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 pseudo-label generation (PLG) with SAM (ablation) | baseline segmenter · COD10K | S-measure: 0.720 -> 0.783 | Improved |
| add pseudo-label storage (PLS_t) (ablation) | baseline + PLG · COD10K | S-measure: 0.783 -> 0.795 | Improved |
| add pseudo-label supervision (PLS_u) (ablation) | baseline + PLG + PLS_t · COD10K | S-measure: 0.795 -> 0.807 | Improved |
| add HGFG module (ablation) | SEE · COD10K | S-measure: 0.785 -> 0.807 | Improved |
| add RK2 aggregation in HGFG (ablation) | SEE · COD10K | S-measure: 0.785 -> 0.807 | Improved |
| use strong augmentation instead of weak augmentation (ablation) | SEE pseudo-label generation · COD10K | S-measure: 0.746 -> 0.807 | Improved |
| add fusion over augmented views on SAM (MARF-S) (ablation) | SEE · COD10K | S-measure: 0.786 -> 0.807 | Improved |
| use SAM2 instead of SAM (ablation) | SEE · COD10K | S-measure: 0.807 vs 0.806 | No clear effect |
| replace WS-SAM with SEE framework | SCOD · COD10K scribble | S-measure: 0.742 -> 0.751 | Improved |
| replace WS-SAM with SEE framework | GenSAM · COD10K scribble | S-measure: 0.810 -> 0.818 | Improved |
| replace WS-SAM with SEE framework | EPS · COD10K 1/8 labeled | S-measure: 0.761 -> 0.766 | Improved |
| increase labeled data from 1/16 to 1/8 | SEE · COD10K SSCOD | S-measure: 0.745 -> 0.786 | Improved |
Closest papers by meaning
- Weakly-Supervised Concealed Object Segmentation with SAM-based Pseudo Labeling and Multi-scale Feature Grouping (NeurIPS 2023 · similarity 0.92)
- When Specialists Meet Generalists: Segmenter-Coordinated Asymmetric Learning for Label-Deficient Concealed Object Segmentation (ECCV 2026 · similarity 0.91)
- UniT: Unified Knowledge Transfer for Any-Shot Object Detection and Segmentation (CVPR 2021 · similarity 0.87)
- Segment Anything without Supervision (NeurIPS 2024 · similarity 0.86)
- RUN: Reversible Unfolding Network for Concealed Object Segmentation (ICML 2025 · similarity 0.85)
- Bridge the Points: Graph-based Few-shot Segment Anything Semantically (NeurIPS 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.
