Research map / Image generation / Autoregressive and latent generation
Thinking Outside the Pool: Active Training Image Creation for Relative Attributes
Aron Yu, Kristen Grauman
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 |
|---|---|---|---|
| Replace half of the real labeled training pairs with human-labeled ATTIC-generated pairs | relative-attribute ranker · UT-Zappos50K, 10 attributes | accuracy | Mixed |
| Replace half of the real labeled training pairs with human-labeled ATTIC-generated pairs | relative-attribute ranker · LFW-10, 8 attributes | accuracy | Mixed |
| Use automatically inferred labels for generated pairs instead of human annotations | ATTIC · UT-Zappos50K, 10 attributes | accuracy | Improved |
| Use automatically inferred labels for generated pairs instead of human annotations | ATTIC · LFW-10, 8 attributes | accuracy | Mixed |
| Use active image generation rather than passive Semantic Jitter | relative-attribute ranker · UT-Zappos50K and LFW active learning curves | accuracy gain over Real: over 3% and 8% | Improved |
| Use low-level geometric and photometric jitter | relative-attribute ranker · UT-Zappos50K and LFW | accuracy | Mixed |
| Use ATTIC rather than Semantic Jitter | relative-attribute ranker · LFW-10, MouthOpen and Young | accuracy | Hurt |
| Use ATTIC rather than Semantic Jitter | relative-attribute ranker · UT-Zappos50K-1, Sporty | accuracy | Hurt |
| Use ATTIC rather than Semantic Jitter | relative-attribute ranker · UT-Zappos50K-2, Open | accuracy | Hurt |
Closest papers by meaning
- Semantic Jitter: Dense Supervision for Visual Comparisons via Synthetic Images (ICCV 2017 · similarity 0.88)
- Learning Attributes Equals Multi-Source Domain Generalization (CVPR 2016 · similarity 0.87)
- Efficient Relative Attribute Learning using Graph Neural Networks (ECCV 2018 · similarity 0.87)
- Attribute2Image: Conditional Image Generation from Visual Attributes (ECCV 2016 · similarity 0.86)
- Attribute Based Interpretable Evaluation Metrics for Generative Models (ICML 2024 · similarity 0.85)
- L-WISE: Boosting Human Visual Category Learning Through Model-Based Image Selection and Enhancement (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.
