Research map / Detection and segmentation / Weakly supervised detection and segmentation
PCL: Proposal Cluster Learning for Weakly Supervised Object Detection
Peng Tang, Xinggang Wang, Song Bai, Wei Shen, Xiang Bai, Wenyu Liu, Alan Yuille
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 |
|---|---|---|---|
| add weighted loss using cluster confidence (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | mAP: 33.6 -> 40.8 | Improved |
| add weighted loss using cluster confidence (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | CorLoc: 51.2 -> 59.6 | Improved |
| use five image scales instead of single scale 600 (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | mAP: 37.4 -> 40.8 | Improved |
| use five image scales instead of single scale 600 (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | CorLoc: 55.5 -> 59.6 | Improved |
| use 3 k-means clusters instead of 2 (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | mAP: 40.2 -> 40.8 | Improved |
| use 3 k-means clusters instead of 4 (ablation) | PCL-OB-G · VOC 2007 test (VGG_M) | mAP: 40.7 -> 40.8 | No clear effect |
| add instance classifier refinement over basic MIL network (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | Improved |
| refine more than 3 times (4 instead of 3) (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | No clear effect |
| graph of top ranking proposals instead of highest scoring proposal for cluster centers (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | Improved |
| treat clusters as bags instead of directly assigning labels (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | Improved |
| online training instead of alternating training (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | Improved |
| set IoU threshold I_t to 0.4 (ablation) | PCL · VOC 2007 test (VGG_M) | mAP | Improved |
Closest papers by meaning
- Multiple Instance Detection Network With Online Instance Classifier Refinement (CVPR 2017 · similarity 0.94)
- Object Instance Mining for Weakly Supervised Object Detection (AAAI 2020 · similarity 0.92)
- Object-Aware Instance Labeling for Weakly Supervised Object Detection (ICCV 2019 · similarity 0.91)
- Instance Mining with Class Feature Banks for Weakly Supervised Object Detection (AAAI 2021 · similarity 0.91)
- C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection (CVPR 2019 · similarity 0.91)
- Instance-Aware, Context-Focused, and Memory-Efficient Weakly Supervised Object Detection (CVPR 2020 · similarity 0.90)
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.
