Research map / Detection and segmentation / Weakly supervised detection and segmentation
Pairwise Similarity Knowledge Transfer for Weakly Supervised Object Localization
Amir Rahimi, Amirreza Shaban, Thalaiyasingam Ajanthan, Richard Hartley, Byron Boots
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 pairwise similarity learning (ablation) | alternating-optimization WSOL · COCO 2017 (17 target classes) | CorLoc@0.5: 39.4 -> 48.3 | Improved |
| add pairwise similarity learning (ablation) | alternating-optimization WSOL · ILSVRC 2013 detection | CorLoc@0.5: 72.8 -> 78.2 | Improved |
| add pairwise similarity learning (ablation) | alternating-optimization WSOL · ILSVRC 2013 detection | CorLoc@0.7: 62.0 -> 65.5 | Improved |
| add pairwise similarity learning (ablation) | alternating-optimization WSOL · ILSVRC 2013 detection | mAP@0.5: 38.1 -> 41.7 | Improved |
| transfer pairwise similarity in warm-up re-localization (ablation) | warm-up re-localization · ILSVRC 2013 detection | CorLoc@0.5: 68.9 -> 73.8 | Improved |
| add re-training and re-localization iterations after warm-up (ablation) | pairwise knowledge transfer · ILSVRC 2013 detection | CorLoc@0.5: 73.8 -> 78.2 | Improved |
| initialize ICM with mini-problem solutions of size K=4 to 64 instead of K=2 (ablation) | ICM re-localization · COCO 2017 warm-up | CorLoc and energy | Improved |
| initialize ICM with random selection instead of objectness (ablation) | ICM re-localization · COCO 2017 warm-up | CorLoc and energy | No clear effect |
| initialize ICM with full-image proposal instead of other schemes (ablation) | ICM re-localization · COCO 2017 warm-up | CorLoc | Hurt |
| use full method instead of Uijlings et al. MIL (Faster-RCNN) | pairwise knowledge transfer · ILSVRC 2013 detection | CorLoc@0.5: 74.2 -> 78.2 | Improved |
Closest papers by meaning
- Evaluation for Weakly Supervised Object Localization: Protocol, Metrics, and Datasets (TPAMI 2022 · similarity 0.93)
- Evaluating Weakly Supervised Object Localization Methods Right (CVPR 2020 · similarity 0.93)
- Rethinking the Route Towards Weakly Supervised Object Localization (CVPR 2020 · similarity 0.93)
- Weakly Supervised Object Localization Using Size Estimates (ECCV 2016 · similarity 0.93)
- Weakly Supervised Object Localization As Domain Adaption (CVPR 2022 · similarity 0.92)
- Boosting Weakly Supervised Object Localization and Segmentation With Domain Adaption (TPAMI 2024 · similarity 0.92)
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.
