Research map / Detection and segmentation / Weakly supervised detection and segmentation
SLV: Spatial Likelihood Voting 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 |
|---|---|---|---|
| Add spatial likelihood voting with re-classification and re-localization (ablation) | Basic PCL · PASCAL VOC 2007 test | mAP: 50.1 -> 52.5 | Helped |
| Add the re-classification branch only (ablation) | Basic PCL · PASCAL VOC 2007 test | mAP: 50.1 -> 51.0 | Helped |
| Add the re-localization branch only (ablation) | Basic PCL · PASCAL VOC 2007 test | mAP: 50.1 -> 51.6 | Helped |
| Train the basic MIL and SLV modules end-to-end instead of separately (ablation) | SLV with re-classification and re-localization · PASCAL VOC 2007 test | mAP: 52.5 -> 53.5 | Helped |
| Add Fast-RCNN retraining on SLV pseudo ground truths (ablation) | SLV end-to-end · PASCAL VOC 2007 test | mAP: 53.5 -> 53.9 | Helped |
| Use spatial likelihood voting rather than selecting the single highest-scoring proposal (ablation) | WSOD pseudo-labeling · PASCAL VOC 2007 test | mAP: 52.1 -> 53.5 | Helped |
| Use spatial likelihood voting rather than selecting the highest-scoring proposal from each cluster (ablation) | WSOD pseudo-labeling · PASCAL VOC 2007 test | mAP: 52.4 -> 53.5 | Helped |
| Use SLV instead of the strongest listed single-model baseline | VGG-based WSOD · PASCAL VOC 2007 test | mAP: 52.9 -> 53.5 | Helped |
| Use SLV instead of Pred Net(VGG) | VGG-based WSOD · PASCAL VOC 2012 test | mAP: 48.4 -> 49.2 | Helped |
| Use SLV instead of Pred Net(VGG) | VGG-based WSOD · PASCAL VOC 2012 trainval | CorLoc: 69.5 -> 69.2 | Hurt |
Closest papers by meaning
- Instance Mining with Class Feature Banks for Weakly Supervised Object Detection (AAAI 2021 · similarity 0.88)
- C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection (CVPR 2019 · similarity 0.87)
- UWSOD: Toward Fully-Supervised-Level Capacity Weakly Supervised Object Detection (NeurIPS 2020 · similarity 0.87)
- W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (CVPR 2018 · similarity 0.86)
- SIOD: Single Instance Annotated per Category per Image for Object Detection (CVPR 2022 · similarity 0.86)
- Weakly Supervised Few-Shot Object Detection with DETR (AAAI 2024 · similarity 0.85)
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.