Research map / Label-efficient and robust learning / Recognition and long-tail
A Deep Cascade Network for Unaligned Face Attribute Classification
Read the paperCheck 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 |
|---|---|---|---|
| Combine learned part-based and whole-image subnet predictions with RSL and ARL | PaW · uCelebA | average classification accuracy: 90.46 -> 91.23; 90.60 -> 91.23 | Helped |
| Add multi-net learning to the localization branch (ablation) | FRL · uCelebA | average classification accuracy: 91.01 -> 91.07 | Helped |
| Add multi-net learning to the localization branch (ablation) | FRL · CUB-200-2011 full image | accuracy: 67.40 -> 71.66 | Helped |
| Add multi-net learning to the localization branch (ablation) | FRL · CUB-200-2011 crop | accuracy: 71.90 -> 76.03 | Helped |
| Use hint-compressed student networks with progressively smaller architectures (ablation) | uCelebA classification model · uCelebA | average classification accuracy: 91.07 -> 91.02; 91.07 -> 90.89; 91.07 -> 90.60 | Hurt |
| Replace learned attribute-specific regions with fixed 4×4 image blocks (ablation) | PaW-style classifier · uCelebA | average classification accuracy: 91.23 -> 90.95 | Hurt |
| Use SNet3 instead of larger compressed architectures (ablation) | uCelebA classification model · uCelebA | parameter count: 19M -> 0.27M | Helped |
| Use SNet3 rather than SOMP-branch-32 or Low Rank | aligned CelebA classifier · aligned CelebA | accuracy: 90.74 -> 90.89; 90.88 -> 90.89 | Helped |
Closest papers by meaning
- Improving Facial Attribute Prediction Using Semantic Segmentation (CVPR 2017 · similarity 0.87)
- Attributes for Improved Attributes: A Multi-Task Network Utilizing Implicit and Explicit Relationships for Facial Attribute Classification (AAAI 2017 · similarity 0.86)
- Partially Shared Multi-Task Convolutional Neural Network With Local Constraint for Face Attribute Learning (CVPR 2018 · similarity 0.85)
- Self-Reinforced Cascaded Regression for Face Alignment (AAAI 2018 · similarity 0.84)
- Attributes Shape the Embedding Space of Face Recognition Models (ICML 2025 · similarity 0.84)
- LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition (CVPR 2024 · similarity 0.83)
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.