Research map / Label-efficient and robust learning / Noisy labels
Learn from All: Erasing Attention Consistency for Noisy Label Facial Expression Recognition
Yuhang Zhang, Chengrui Wang, Xu Ling, Weihong Deng
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 flip attention consistency (ablation) | EAC (erasing + imbalanced framework) · RAF-DB (30% label noise) | accuracy (%): 78.10 -> 84.42 | Improved |
| add the imbalanced framework (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 78.29 -> 84.42 | Improved |
| add random erasing (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 76.26 -> 84.42 | Improved |
| use flip instead of rotate as the spatial transform (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 78.23 -> 84.42 | Improved |
| use flip instead of scale as the spatial transform (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 82.01 -> 84.42 | Improved |
| use erasing instead of blur (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 76.92 -> 84.42 | Improved |
| use erasing instead of AutoAugment (ablation) | EAC · RAF-DB (30% label noise) | accuracy (%): 82.40 -> 84.42 | Improved |
| add EAC to MobileNet | MobileNet · RAF-DB (30% label noise) | accuracy (%): 62.48 -> 79.82 | Improved |
| add EAC to ResNet-50 | ResNet-50 · RAF-DB (30% label noise) | accuracy (%): 71.67 -> 85.27 | Improved |
| use EAC instead of RUL | ResNet-18 noisy FER · RAF-DB (30% label noise) | accuracy (%): 82.06 -> 84.42 | Improved |
| use EAC instead of baseline | ResNet-18 on Tiny-ImageNet · Tiny-ImageNet (30% label noise) | top-1 accuracy (%): 41.32 -> 63.51 | Improved |
| use EAC instead of KTN on clean FERPlus | ResNet-18 · FERPlus (clean) | accuracy (%): 90.49 -> 89.64 | Hurt |
Closest papers by meaning
- CLEX: Complementary Label Exchange Learning for Noisy Facial Expression Recognition (CVPR 2026 · similarity 0.91)
- Navigating Label Ambiguity for Facial Expression Recognition in the Wild (AAAI 2025 · similarity 0.90)
- Facial Expression Recognition with Inconsistently Annotated Datasets (ECCV 2018 · similarity 0.88)
- RepFace: Refining Closed-Set Noise with Progressive Label Correction for Face Recognition (AAAI 2025 · similarity 0.87)
- Open-Set Facial Expression Recognition (AAAI 2024 · similarity 0.87)
- Facial-R1: Aligning Reasoning and Recognition for Facial Emotion Analysis (AAAI 2026 · 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.
