Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly Supervised Localization Using Deep Feature Maps
Archith John Bency, Heesung Kwon, Hyungtae Lee, S. Karthikeyan, B. S. Manjunath
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 |
|---|---|---|---|
| use VGG-16 instead of AlexNet as classifier | Proposed · PASCAL VOC 2012 val | localization mAP: 77.1 -> 79.7 | Improved |
| use VGG-16 instead of AlexNet as classifier | Proposed · PASCAL VOC 2012 val | classification mAP: 79.8 -> 86.5 | Improved |
| add label co-occurrence re-scoring (ablation) | Proposed (VGG-16) · PASCAL VOC 2012 val | classification mAP: +1.2 | Improved |
| add label co-occurrence re-scoring (ablation) | Proposed (VGG-16) · PASCAL VOC 2012 val | localization mAP: +0.8 | Improved |
| weakly supervised feature-map beam search instead of Oquab et al. | Proposed · PASCAL VOC 2012 val | localization mAP: 74.5 -> 79.7 | Improved |
| weakly supervised feature-map beam search instead of Oquab et al. | Proposed · MS COCO val | localization mAP: 41.2 -> 49.2 | Improved |
| weak supervision instead of box-supervised RCNN | Proposed · PASCAL VOC 2012 val | localization mAP: 74.8 -> 79.7 | Improved |
| proposed vs Bilen et al. weakly supervised detector | Proposed · PASCAL VOC 2007 test | detection mAP: 27.7 -> 25.7 | No clear effect |
| proposed vs Oquab et al. with Selective Search | Proposed · PASCAL VOC 2012 val (detection) | detection mAP: 11.7 -> 26.5 | Improved |
| beam search over feature-map windows instead of region proposals | Proposed · Localization time | seconds: 20 -> 2.6 | Improved |
Closest papers by meaning
- Weakly Supervised Object Localization and Detection: A Survey (TPAMI 2021 · similarity 0.93)
- ContextLocNet: Context-Aware Deep Network Models for Weakly Supervised Localization (ECCV 2016 · similarity 0.91)
- Soft Proposal Networks for Weakly Supervised Object Localization (ICCV 2017 · similarity 0.91)
- Bridging the Gap Between Classification and Localization for Weakly Supervised Object Localization (CVPR 2022 · similarity 0.91)
- Two-Phase Learning for Weakly Supervised Object Localization (ICCV 2017 · similarity 0.90)
- Inter-Image Communication for Weakly Supervised Localization (ECCV 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.
