Research map / Detection and segmentation / Open-set and few-shot detection
Multiple Instance Active Learning for Object Detection
Check 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 IUL to random-selection training (ablation) | RetinaNet active detection · PASCAL VOC, 20.0% labeled | mAP: 67.09% vs. 70.06% | Helped |
| add IUL to active image classification (ablation) | ResNet-18 · CIFAR-10, 2.0% labeled | accuracy: 58.07% vs. 51.01% | Helped |
| add IUR using image-classification-score weights (ablation) | MI-AOD with IUL · PASCAL VOC, 20.0% labeled | mAP: 68.48% vs. 67.20% | Helped |
| add IUR using image-classification-score weights (ablation) | MI-AOD with IUL · PASCAL VOC, 100.0% labeled | mAP: 78.37% vs. 77.28% | Helped |
| use image-classification scores rather than uniform weights for IUR (ablation) | MI-AOD · PASCAL VOC, first three active-learning cycles | mAP: 5.04%~17.09% | Helped |
| use image-classification weights rather than first-classifier outputs (ablation) | IUR · PASCAL VOC, labeled proportions 5.0%–20.0% | mAP: 31.67 -> 47.18; 67.30 -> 68.48 | Helped |
| apply image-classification-score weights on all data rather than only the labeled set (ablation) | IUR · PASCAL VOC, 20.0% labeled | mAP: 68.32 -> 68.48 | Helped |
| use image-classification-score weights rather than uniform weights on the labeled set (ablation) | IUR · PASCAL VOC, 5.0% labeled | mAP: 42.52 -> 47.18 | Helped |
| use λ=0.5 rather than λ=1 (ablation) | MI-AOD · PASCAL VOC, 20.0% labeled | mAP: 71.62 -> 72.27 | Helped |
| select top 10k instances rather than all instances (ablation) | MI-AOD · PASCAL VOC, 20.0% labeled | mAP: 72.03 -> 72.27 | Helped |
| select top 10k instances rather than 100 instances (ablation) | MI-AOD · PASCAL VOC, 20.0% labeled | mAP: 71.47 -> 72.27 | Helped |
| use mean instance uncertainty rather than maximum instance uncertainty (ablation) | IUL + IUR · PASCAL VOC, 20.0% labeled | mAP: 70.33 -> 72.03 | Helped |
Closest papers by meaning
- Portable Active Learning for Object Detection (CVPR 2026 · similarity 0.91)
- Active Learning for Object Detection with Evidential Deep Learning and Hierarchical Uncertainty Aggregation (ICLR 2023 · similarity 0.88)
- Not All Labels Are Equal: Rationalizing the Labeling Costs for Training Object Detection (CVPR 2022 · similarity 0.87)
- Plug and Play Active Learning for Object Detection (CVPR 2024 · similarity 0.87)
- Towards Human-Machine Cooperation: Self-Supervised Sample Mining for Object Detection (CVPR 2018 · similarity 0.87)
- Multiple Anchor Learning for Visual Object Detection (CVPR 2020 · 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.