Research map / Detection and segmentation / Weakly supervised detection and segmentation
Boosting Weakly Supervised Object Detection via Learning Bounding Box Adjusters
Bowen Dong, Zitong Huang, Yuelin Guo, Qilong Wang, Zhenxing Niu, Wangmeng Zuo
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 |
|---|---|---|---|
| use R50-C4 backbone for the adjuster instead of VGG-16 (ablation) | LBBA with WSDDN · VOC 2007 | mAP / CorLoc: 50.2 -> 52.7 mAP; 67.7 -> 70.3 CorLoc | Improved |
| add LBBA to WSDDN (ablation) | WSDDN · VOC 2007 | mAP: 46.6 -> 52.7 | Improved |
| add LBBA to OICR+REG (ablation) | OICR · VOC 2007 | mAP: 48.6 -> 55.1 | Improved |
| add LBBA to OICR+REG with top-p% pseudo-label mining (ablation) | OICR+ · VOC 2007 | mAP: 51.4 -> 55.8 | Improved |
| add LBBA to WSDDN (ablation) | WSDDN · VOC 2007 | CorLoc: 64.7 -> 70.3 | Improved |
| add LBBA to OICR+REG with top-p% pseudo-label mining (ablation) | OICR+ · VOC 2007 | CorLoc: 64.9 -> 71.6 | Improved |
| increase multi-stage iterations T from 0 to 3 (ablation) | LBBA with OICR+ · VOC 2007 | mAP: 53.1 -> 55.8 | Improved |
| increase multi-stage iterations T from 0 to 3 (ablation) | LBBA adjuster · COCO-60 | adjuster mAP: 29.1 -> 30.9 | Improved |
| increase T beyond 2 (ablation) | LBBA with OICR+ · VOC 2007 | mAP: 55.7 -> 55.8 | No clear effect |
| use learnable auxiliary WSOD network instead of MCG proposals (ablation) | LBBA · VOC 2007 | mAP: 54.3 -> 55.8 | Improved |
| use learnable auxiliary WSOD network instead of MCG proposals (ablation) | LBBA adjuster · COCO-60 | adjuster mAP: 29.6 -> 30.9 | Improved |
| add classification score masking strategy (ablation) | LBBA · COCO-20 | mAP: 13.0 -> 13.7 | Improved |
Closest papers by meaning
- GradingNet: Towards Providing Reliable Supervisions for Weakly Supervised Object Detection by Grading the Box Candidates (AAAI 2021 · similarity 0.90)
- W2N: Switching from Weak Supervision to Noisy Supervision for Object Detection (ECCV 2022 · similarity 0.90)
- Towards Precise End-to-End Weakly Supervised Object Detection Network (ICCV 2019 · similarity 0.89)
- Bounding Box Regression With Uncertainty for Accurate Object Detection (CVPR 2019 · similarity 0.89)
- Weakly Supervised Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.87)
- Weakly Supervised Few-Shot Object Detection with DETR (AAAI 2024 · similarity 0.87)
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.
