Research map / Detection and segmentation / Weakly supervised detection and segmentation
CamoTeacher: Dual-Rotation Consistency Learning for Semi-Supervised Camouflaged Object Detection
xunfa lai, Zhiyu Yang, Jie Hu, ShengChuan Zhang, Liujuan Cao, Guannan Jiang, Songan Zhang, zhiyu wang, Rongrong Ji
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 |
|---|---|---|---|
| apply uniform Mean Teacher pseudo-labels instead of supervised-only | SINet · NC4K (1% labeled) | S-measure: 0.618 -> 0.541 | Hurt |
| replace Mean Teacher with CamoTeacher (DRCL) | SINet · COD10K (1% labeled) | S-measure: 0.546 -> 0.699 | Improved |
| add pixel-wise consistency weight omega_pc to BCE loss (ablation) | CamoTeacher · COD10K (5% labeled) | weighted F-measure: 0.370 -> 0.501 | Improved |
| add instance-wise consistency weight omega_ic to IoU loss (ablation) | CamoTeacher · COD10K (5% labeled) | weighted F-measure: 0.572 -> 0.585 | Improved |
| combine PCL and ICL weighted losses (ablation) | CamoTeacher · COD10K (5% labeled) | weighted F-measure: 0.370 -> 0.586 | Improved |
| combine pixel inconsistency and distance-from-0.5 terms in pixel weight (ablation) | CamoTeacher PCL · COD10K (5% labeled) | weighted F-measure: 0.485 -> 0.501 | Improved |
| multiply pixel inconsistency weight by raw prediction instead of uncertainty term (ablation) | CamoTeacher PCL · COD10K (5% labeled) | weighted F-measure: 0.481 -> 0.405 | Hurt |
| apply CamoTeacher to FSPNet with 10% labels | FSPNet · COD10K (10% labeled) | S-measure: 0.789 -> 0.817 | Improved |
| add unlabeled CAMO and NC4K data to 100% labeled training | CamoTeacher SINet · COD10K | S-measure: 0.771 -> 0.787 | Improved |
| set alpha to 1/4 rather than 2 (ablation) | CamoTeacher PCL · COD10K (5% labeled) | weighted F-measure: 0.565 -> 0.579 | Improved |
Closest papers by meaning
- Learning Camouflaged Object Detection from Noisy Pseudo Label (ECCV 2024 · similarity 0.89)
- SAM-COD: SAM-guided Unified Framework for Weakly-Supervised Camouflaged Object Detection (ECCV 2024 · similarity 0.88)
- Multi-clue Consistency Learning to Bridge Gaps Between General and Oriented Object in Semi-supervised Detection (AAAI 2025 · similarity 0.88)
- Just a Hint: Point-Supervised Camouflaged Object Detection (ECCV 2024 · similarity 0.87)
- Beyond Weak Supervision: MLLMs-Guided Graded Knowledge Distillation for Unsupervised Camouflaged Object Detection (CVPR 2026 · similarity 0.87)
- UCOD-DPL: Unsupervised Camouflaged Object Detection via Dynamic Pseudo-label Learning (CVPR 2025 · 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.
