Research map / Label-efficient and robust learning / Recognition and long-tail
CQA-Face: Contrastive Quality-Aware Attentions for Face Recognition
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 |
|---|---|---|---|
| Add CAL and QAN to the face recognition model | CQA-Face · Masked2Normal (M2N) | TAR at FAR=0.1: 33.43 -> 59.40 | Helped |
| Add CAL and QAN to the face recognition model | CQA-Face · Masked2Masked (M2M) | TAR at FAR=0.001: 2.95 -> 17.75 | Helped |
| Add CAL and QAN to the face recognition model | CQA-Face · MLFW | accuracy: 74.61 -> 92.78 | Helped |
| Add CAL and QAN to the face recognition model | CQA-Face · IJB-A quality, FAR=0.001 | TAR: 65.7 -> 86.4 | Helped |
| Add CAL and QAN to the face recognition model | CQA-Face · CPLFW | accuracy: 92.63 -> 93.00 | Helped |
| Add CAL and QAN to the face recognition model | CQA-Face · LFW | accuracy: 99.83 -> 99.83 | No clear effect |
| Add CAL to the baseline attention model (ablation) | CQA-Face · CALFW | accuracy: 92.55 -> 92.90 | Helped |
| Add CAL to the baseline attention model (ablation) | CQA-Face · CPLFW | accuracy: 89.55 -> 89.85 | Helped |
| Add QAN to the model with CAL (ablation) | CQA-Face · CPLFW | accuracy: 89.85 -> 90.20 | Helped |
| Replace CAL with the LD diverse-learning loss (ablation) | CQA-Face · LFW | accuracy: 99.47 -> 99.33 | Hurt |
| Replace CAL with dropout (ablation) | CQA-Face · CPLFW | accuracy: 89.85 -> 89.52 | Hurt |
| Use 2 local branches instead of 4 (ablation) | CQA-Face · CPLFW | accuracy: 90.20 -> 90.02 | Hurt |
Closest papers by meaning
- CR-FIQA: Face Image Quality Assessment by Learning Sample Relative Classifiability (CVPR 2023 · similarity 0.86)
- SDD-FIQA: Unsupervised Face Image Quality Assessment With Similarity Distribution Distance (CVPR 2021 · similarity 0.86)
- CLIB-FIQA: Face Image Quality Assessment with Confidence Calibration (CVPR 2024 · similarity 0.85)
- Integrating Global Context Contrast and Local Sensitivity for Blind Image Quality Assessment (ICML 2024 · similarity 0.85)
- FaceBench: A Multi-View Multi-Level Facial Attribute VQA Dataset for Benchmarking Face Perception MLLMs (CVPR 2025 · similarity 0.84)
- DSL-FIQA: Assessing Facial Image Quality via Dual-Set Degradation Learning and Landmark-Guided Transformer (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.