Research map / Detection and segmentation / Weakly supervised detection and segmentation
End-to-End Localization and Ranking for Relative Attributes
Krishna Kumar Singh, Yong Jae Lee
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 |
|---|---|---|---|
| add the STN-localized patch to a global-image-only ranker (ablation) | Relative attribute ranker · LFW-10 (mean) | ranking accuracy (%): 82.51 -> 86.91 | Improved |
| add the global image input to an STN-output-only ranker (ablation) | Relative attribute ranker · LFW-10 (mean) | ranking accuracy (%): 84.63 -> 86.91 | Improved |
| use STN output only instead of global image only (ablation) | Relative attribute ranker · LFW-10 (mean) | ranking accuracy (%): 82.51 -> 84.63 | Improved |
| global image only is better than STN-only on bald-head (ablation) | Relative attribute ranker · LFW-10 (bald-head) | ranking accuracy (%): 84.31 -> 78.10 | Hurt |
| end-to-end joint training instead of the pipeline of Xiao and Lee | Relative attribute ranking · LFW-10 (mean) | ranking accuracy (%): 84.66 -> 86.91 | Improved |
| end-to-end joint training instead of the pipeline of Xiao and Lee | Relative attribute ranking · UT-Zap50K-1 (mean) | ranking accuracy (%): 95.47 -> 95.78 | No clear effect |
| end-to-end joint training instead of the pipeline of Xiao and Lee | Relative attribute ranking · Shoes-with-Attribute (cross-dataset) | ranking accuracy (%): 83.58 -> 88.46 | Improved |
| end-to-end STN ranker instead of local learning of Yu and Grauman | Relative attribute ranking · OSR (mean) | ranking accuracy (%): 92.37 -> 97.02 | Improved |
| end-to-end network instead of the multi-stage pipeline | Relative attribute ranking · test time per image | seconds: 1.1 -> 0.011 | Improved |
Closest papers by meaning
- Region Similarity Representation Learning (ICCV 2021 · similarity 0.85)
- Patch Reordering: A NovelWay to Achieve Rotation and Translation Invariance in Convolutional Neural Networks (AAAI 2017 · similarity 0.85)
- Unsupervised Learning of Discriminative Attributes and Visual Representations (CVPR 2016 · similarity 0.85)
- Soft Contrastive Learning for Visual Localization (NeurIPS 2020 · similarity 0.85)
- On the rankability of visual embeddings (NeurIPS 2025 · similarity 0.85)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 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.
