Research map / Detection and segmentation / Weakly supervised detection and segmentation
ContextLocNet: Context-Aware Deep Network Models for Weakly Supervised Localization
Vadim Kantorov, Maxime Oquab, Minsu Cho, Ivan Laptev
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 |
|---|---|---|---|
| replace the additive model with the contrastive S model (shared FC layers, frame pooling) (ablation) | context-aware WSL · VOC 2007 test | mAP: 33.3 -> 36.3 | Improved |
| replace the additive model with the contrastive S model (ablation) | context-aware WSL · VOC 2007 trainval | CorLoc: 52.8 -> 55.1 | Improved |
| add frame pooling (symmetric) to the contrastive model instead of ROI pooling for context (ablation) | contrastive model · VOC 2007 test | mAP: 32.2 -> 36.3 | Improved |
| add frame pooling (symmetric) to the contrastive model instead of ROI pooling for context (ablation) | contrastive model · VOC 2007 trainval | CorLoc: 50.2 -> 55.1 | Improved |
| add contrastive S context scoring on top of the WSDDN-style baseline (ablation) | WSDDN-SSW-S* reimplementation · VOC 2007 test | mAP: 30.5 -> 36.3 | Improved |
| add contrastive S context scoring on top of the WSDDN-style baseline (ablation) | WSDDN-SSW-S* reimplementation · VOC 2007 trainval | CorLoc: 50.0 -> 55.1 | Improved |
| add additive context scoring to the WSDDN-style baseline (ablation) | WSDDN-SSW-S* reimplementation · VOC 2007 test | mAP: 30.5 -> 33.3 | Improved |
| add asymmetric contrastive context scoring to the WSDDN-style baseline (ablation) | WSDDN-SSW-S* reimplementation · VOC 2007 test | mAP: 30.5 -> 32.2 | Improved |
| add asymmetric contrastive context scoring to the WSDDN-style baseline (ablation) | WSDDN-SSW-S* reimplementation · VOC 2007 trainval | CorLoc: 50.0 -> 50.2 | No clear effect |
| train additive and contrastive models jointly by adding their outputs (ablation) | context models · VOC 2007 test | mAP | No clear effect |
| add other types of region pooling as input to the localization branch (ablation) | context models · VOC 2007 test | mAP | No clear effect |
| share weights or replace context pooling with frame pooling in the additive model (ablation) | additive model · VOC 2007 test | mAP | Hurt |
Closest papers by meaning
- Weakly Supervised Localization Using Deep Feature Maps (ECCV 2016 · similarity 0.91)
- Weakly Supervised Object Localization and Detection: A Survey (TPAMI 2021 · similarity 0.90)
- Soft Proposal Networks for Weakly Supervised Object Localization (ICCV 2017 · similarity 0.89)
- Weakly Supervised Object Localization via Transformer with Implicit Spatial Calibration (ECCV 2022 · similarity 0.89)
- Learning to Localize Objects with Noisy Labeled Instances (AAAI 2019 · similarity 0.89)
- Weakly Supervised Object Localization With Progressive Domain Adaptation (CVPR 2016 · 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.
