Research map / Detection and segmentation / Weakly supervised detection and segmentation
Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection
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 |
|---|---|---|---|
| transfer tracked video boxes to weakly-labeled images and train on the resulting pseudo boxes | weakly-supervised object detection · VOC 2007 test | mAP: 34.9 -> 41.9 | Helped |
| transfer tracked video boxes to weakly-labeled images and train on the resulting pseudo boxes | weakly-supervised object detection · VOC 2010 test | mAP: 32.1 -> 40.1 | Helped |
| add LSVM updating to discover missed pseudo boxes and correct existing ones (ablation) | R-CNN trained with initial pseudo GT · VOC 2007 test | mAP: 34.4 -> 36.9 | Helped |
| add bounding-box regression (ablation) | R-CNN trained with updated pseudo GT · VOC 2007 test | mAP: 36.9 -> 39.7 | Helped |
| fine-tune CNN features using pseudo GT boxes (ablation) | R-CNN with updated pseudo GT and bounding-box regression · VOC 2007 test | mAP: 39.7 -> 41.9 | Helped |
| apply LSVM updating to detect missed boxes and update existing boxes (ablation) | initial pseudo GT boxes · VOC 2007 train+val | CorLoc: 40.9 -> 51.7 | Helped |
| exclude training images in which no initial pseudo GT was found (ablation) | initial pseudo GT evaluation · VOC 2007 train+val | CorLoc: 40.9 -> 51.1 | Helped |
Closest papers by meaning
- W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (CVPR 2018 · similarity 0.92)
- You Reap What You Sow: Using Videos to Generate High Precision Object Proposals for Weakly-Supervised Object Detection (CVPR 2019 · similarity 0.90)
- Activity Driven Weakly Supervised Object Detection (CVPR 2019 · similarity 0.89)
- Weakly Supervised Object Boundaries (CVPR 2016 · similarity 0.89)
- We Don't Need No Bounding-Boxes: Training Object Class Detectors Using Only Human Verification (CVPR 2016 · similarity 0.89)
- Relational Matching for Weakly Semi-Supervised Oriented Object Detection (CVPR 2024 · similarity 0.87)
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.