Research map / Detection and segmentation / Weakly supervised detection and segmentation
Deep Self-Taught Learning for Weakly Supervised Object Localization
Check 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 dense subgraph discovery to HCP seed selection (ablation) | HCP · PASCAL 2007 trainval | CorLoc: 34.1 -> 37.9 | Helped |
| Add dense subgraph discovery to HCP seed selection (ablation) | HCP · PASCAL 2007 test | mAP: 26.7 -> 29.6 | Helped |
| Add dense subgraph discovery to HCP seed selection (ablation) | HCP · PASCAL 2012 val | mAP: 24.7 -> 28.3 | Helped |
| Add one round of online supportive sample harvesting using relative score improvement (ablation) | HCP+DSD · PASCAL 2007 trainval | CorLoc: 37.9 -> 50.2 | Helped |
| Add one round of online supportive sample harvesting (ablation) | HCP+DSD · PASCAL 2007 test | mAP: 29.6 -> 38.8 | Helped |
| Perform additional OSSH rounds after the first round (ablation) | HCP+DSD+OSSH1 · PASCAL 2007 trainval | CorLoc: 50.2 -> 56.1 | Helped |
| Add negative rejection after OSSH3 (ablation) | HCP+DSD+OSSH3 · PASCAL 2007 test | mAP: 40.8 -> 41.7 | Helped |
| Add negative rejection after OSSH3 (ablation) | HCP+DSD+OSSH3 · PASCAL 2012 val | mAP: 38.5 -> 39.0 | Helped |
| Use relative score improvement rather than absolute CNN scores to select supportive samples (ablation) | OSSH · PASCAL 2007 trainval | CorLoc: 53.2 -> 56.1 | Helped |
| Train with PASCAL 2007 and 2012 trainval instead of PASCAL 2007 trainval alone | HCP+DSD+OSSH3+NR · PASCAL 2007 test | mAP: 41.7 -> 43.7 | Helped |
| Train with PASCAL 2007 and 2012 trainval instead of PASCAL 2012 trainval alone | HCP+DSD+OSSH3+NR · PASCAL 2012 test | mAP: 38.3 -> 39.4 | Helped |
Closest papers by meaning
- Learning to Localize Objects with Noisy Labeled Instances (AAAI 2019 · similarity 0.89)
- Evaluating Weakly Supervised Object Localization Methods Right (CVPR 2020 · similarity 0.88)
- Unveiling the Potential of Structure Preserving for Weakly Supervised Object Localization (CVPR 2021 · similarity 0.88)
- Weakly Supervised Object Localization With Progressive Domain Adaptation (CVPR 2016 · similarity 0.88)
- Self-Supervised Object Localization with Joint Graph Partition (AAAI 2022 · similarity 0.88)
- Rethinking the Route Towards Weakly Supervised Object Localization (CVPR 2020 · 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.