Research map / Detection and segmentation / Weakly supervised detection and segmentation
UWSOD: Toward Fully-Supervised-Level Capacity Weakly Supervised 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 step-wise bounding-box fine-tuning (nf=4) (ablation) | UWSOD with SSOPG · PASCAL VOC 2007 test | mAP (%): 32.3 -> 42.1 | Helped |
| add step-wise bounding-box fine-tuning (nf=4) (ablation) | UWSOD with SSOPG · PASCAL VOC 2007 test | CorLoc (%): 50.9 -> 60.8 | Helped |
| add multi-rate resampling pyramid (nm=3) (ablation) | UWSOD with SSOPG+SWBBFT · PASCAL VOC 2007 test | mAP (%): 42.1 -> 44.0 | Helped |
| add multi-rate resampling pyramid (nm=3) (ablation) | UWSOD with SSOPG+SWBBFT · PASCAL VOC 2007 test | CorLoc (%): 60.8 -> 62.6 | Helped |
| replace external-proposal-free baseline with self-supervised proposal generator (SSOPG) (ablation) | WSDDN reimplementation · PASCAL VOC 2007 test | mAP (%): 36.7 -> 32.3 | Hurt |
| increase SSOPG proposals from 1024 to 2048 (ablation) | SSOPG · PASCAL VOC 2007 test | mAP (%): 32.1 -> 32.3 | No clear effect |
| increase SSOPG proposals from 2048 to 4096 (ablation) | SSOPG · PASCAL VOC 2007 test | mAP (%): 32.3 -> 32.1 | No clear effect |
| increase SWBBFT branches from 4 to 5 (ablation) | SWBBFT · PASCAL VOC 2007 test | mAP (%): 42.1 -> 42.0 | No clear effect |
| reduce SWBBFT branches from 4 to 3 (ablation) | SWBBFT · PASCAL VOC 2007 test | mAP (%): 42.1 -> 41.8 | No clear effect |
| increase MRRP rates from nm=3 to nm=4 (ablation) | MRRP · PASCAL VOC 2007 test | mAP (%): 44.0 -> 43.6 | No clear effect |
| map proposals to own pyramid level instead of averaging feature maps (ablation) | MRRP nm=3 · PASCAL VOC 2007 test | CorLoc (%): 62.6 -> 63.0 | No clear effect |
| add class-agnostic GT boxes for regression supervision (ablation) | UWSOD · PASCAL VOC 2007 test | mAP (%): 44.0 -> 67.7 | Helped |
Closest papers by meaning
- Salvage of Supervision in Weakly Supervised Object Detection (CVPR 2022 · similarity 0.91)
- GradingNet: Towards Providing Reliable Supervisions for Weakly Supervised Object Detection by Grading the Box Candidates (AAAI 2021 · similarity 0.90)
- Object Instance Mining for Weakly Supervised Object Detection (AAAI 2020 · similarity 0.90)
- Weakly Supervised Few-Shot Object Detection with DETR (AAAI 2024 · similarity 0.90)
- Weakly Supervised Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.89)
- Semi-supervised Open-World Object Detection (AAAI 2024 · similarity 0.88)
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.