Research map / Detection and segmentation / Open-set and few-shot detection
Mining Object Parts From CNNs via Active Question-Answering
Quanshi Zhang, Ruiming Cao, Ying Nian Wu, Song-Chun Zhu
Check 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 active QA to mine and update AOG part knowledge rather than learning the AOG without QA | AOG part localization · CUB200-2011 | normalized distance: 0.1084 -> 0.0626 | Improved |
| Use active-QA AOG localization with 10 annotations instead of Fast-RCNN (2 fts) with 30 annotations | Part localization · PASCAL VOC Part, head | average normalized distance: 0.161 -> 0.148 | Improved |
| Use active-QA AOG localization with 10 annotations instead of Fast-RCNN (1 ft) with 30 annotations | Part localization · PASCAL VOC Part, neck | average normalized distance: 0.258 -> 0.152 | Improved |
| Use active-QA AOG localization with 10 annotations instead of Fast-RCNN (2 fts) with 30 annotations | Part localization · PASCAL VOC Part, nose/muzzle/beak | average normalized distance: 0.297 -> 0.164 | Improved |
| Add Fast-RCNN localization refinement to the AOG method | Ours · PASCAL VOC Part | PCP: 33.0; 47.2; 50.5 | Mixed |
| Increase the number of part annotations from 10 to 20 | Ours · CUB200-2011 | normalized distance: 0.0626 -> 0.0434 | Improved |
| Increase the number of part annotations from 10 to 30 | Ours · ILSVRC 2013 DET Animal-Part | average normalized distance: 0.1126 -> 0.0909 | Improved |
Closest papers by meaning
- Mining Interpretable AOG Representations From Convolutional Networks via Active Question Answering (TPAMI 2020 · similarity 0.96)
- Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning (AAAI 2017 · similarity 0.88)
- Do Semantic Parts Emerge in Convolutional Neural Networks? (IJCV 2017 · similarity 0.86)
- Interpreting CNN Knowledge via an Explanatory Graph (AAAI 2018 · similarity 0.86)
- Extraction of an Explanatory Graph to Interpret a CNN (TPAMI 2020 · similarity 0.86)
- Exploring the Effectiveness of Object-Centric Representations in Visual Question Answering: Comparative Insights with Foundation Models (ICLR 2025 · similarity 0.84)
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.
