Research map / Detection and segmentation / Open-set and few-shot detection
Towards Human-Machine Cooperation: Self-Supervised Sample Mining for Object Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 cross-image validated SSM and active learning, with 200% annotation and 1000% pseudo-labeling | RFCN · VOC 2007 test | mAP: 79.1$\pm$0.4 -> 80.6$\pm$0.2 | Helped |
| Add cross-image validated SSM and active learning, with 100% annotation and 800% pseudo-labeling | RFCN · VOC 2012 test | mAP: 76.8$\pm$0.4 -> 78.1$\pm$0.2 | Helped |
| Use SSM instead of SPL for pseudo-label mining (ablation) | RFCN · VOC 2007 test | mAP: 74.7$\pm$0.6 -> 76.7$\pm$0.3 | Helped |
| Use SSM instead of SPL for pseudo-label mining (ablation) | RFCN · VOC 2012 test | mAP: 70.9$\pm$0.5 -> 72.1$\pm$0.3 | Helped |
| Add SSM without active learning (ablation) | RFCN · VOC 2007 test | mAP: 73.9 -> 76.7$\pm$0.3 | Helped |
| Add SSM without active learning (ablation) | RFCN · VOC 2012 test | mAP: 69.1 -> 72.1$\pm$0.3 | Helped |
| Add active learning without SSM (ablation) | RFCN · VOC 2007 test | mAP: 73.9 -> 77.5$\pm$0.2 | Helped |
| Add active learning without SSM (ablation) | RFCN · VOC 2012 test | mAP: 69.1 -> 74.7$\pm$0.2 | Helped |
| Add SSM without active learning to a detector initialized with 5% annotations (ablation) | RFCN · VOC 2007 test | mAP: 49.4$\pm$1.4 -> 51.6$\pm$0.3 | Helped |
| Add SSM without active learning to a detector initialized with 10% annotations (ablation) | RFCN · VOC 2007 test | mAP: 56.8$\pm$1.3 -> 59.4$\pm$0.2 | Helped |
| Use SSM and active learning instead of random annotation selection at 5% initialization and 30% added annotations | RFCN · VOC 2007 test | mAP: 53.3$\pm$1.1 -> 55.1$\pm$0.1 | Helped |
| Use SSM and active learning instead of random annotation selection at 10% initialization and 30% added annotations | RFCN · VOC 2007 test | mAP: 60.4$\pm$1.5 -> 62.9$\pm$0.2 | Helped |
Closest papers by meaning
- Portable Active Learning for Object Detection (CVPR 2026 · similarity 0.89)
- Not All Labels Are Equal: Rationalizing the Labeling Costs for Training Object Detection (CVPR 2022 · similarity 0.89)
- Training Region-Based Object Detectors With Online Hard Example Mining (CVPR 2016 · similarity 0.88)
- Plug and Play Active Learning for Object Detection (CVPR 2024 · similarity 0.88)
- LEOD: Label-Efficient Object Detection for Event Cameras (CVPR 2024 · similarity 0.86)
- SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection (CVPR 2025 · 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.