Research map / Detection and segmentation / Weakly supervised detection and segmentation
Noise-Aware Fully Webly Supervised Object Detection
Yunhang Shen, Rongrong Ji, Zhiwei Chen, Xiaopeng Hong, Feng Zheng, Jianzhuang Liu, Mingliang Xu, Qi Tian
Check 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 the noise-aware framework rather than direct WSDDN training on web images | VGG16 detector · PASCAL VOC 2007 test | mAP: 27.6 -> 35.1 | Improved |
| Add a residual detection head (ablation) | Webly supervised VGG16 detector · PASCAL VOC 2007 test | mAP: 27.6 -> 29.8 | Improved |
| Use original entropy weighting for background labels (ablation) | Residual detection model · PASCAL VOC 2007 test | mAP: 27.6 -> 30.7 | Improved |
| Use spatially-sensitive entropy rather than the original entropy weighting (ablation) | Residual detection model · PASCAL VOC 2007 test | mAP: 30.7 -> 33.4 | Improved |
| Apply SSE to foreground as well as background categories (ablation) | Residual detection model with SSE · PASCAL VOC 2007 test | mAP: 33.4 -> 33.6 | Improved |
| Add bagging-mixup using two images per bag (ablation) | Residual detection model with SSE · PASCAL VOC 2007 test | mAP: 33.4 -> 35.1 | Improved |
| Increase bagging-mixup bag size from 2 to 3 images (ablation) | Residual detection model with SSE and bagging-mixup · PASCAL VOC 2007 test | mAP: 35.1 -> 35.2 | No clear effect |
| Increase bagging-mixup bag size from 2 to 4 images (ablation) | Residual detection model with SSE and bagging-mixup · PASCAL VOC 2007 test | mAP: 35.1 -> 35.1 | No clear effect |
| Train on Flickr-COCO with the noise-aware method rather than WSDDN | VGG16 detector · MS COCO minival | AP0.5: 7.0 -> 10.6 | Improved |
Closest papers by meaning
- Exploring Bottom-Up and Top-Down Cues With Attentive Learning for Webly Supervised Object Detection (CVPR 2020 · similarity 0.89)
- Improving Object Detection with Selective Self-Supervised Self-Training (ECCV 2020 · similarity 0.88)
- W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (CVPR 2018 · similarity 0.88)
- W2N: Switching from Weak Supervision to Noisy Supervision for Object Detection (ECCV 2022 · similarity 0.88)
- An accurate detection is not all you need to combat label noise in web-noisy datasets (ECCV 2024 · similarity 0.87)
- Search and Detect: Training-Free Long Tail Object Detection via Web-Image Retrieval (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.
