Research map / Label-efficient and robust learning / Recognition and long-tail
Adversarial Pose Regression Network for Pose-Invariant Face Recognitions
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 APRN to CASIA-Net | CASIA-Net · CFP-FP | accuracy: 90.46% to 94.61% | Helped |
| Add APRN to CASIA-Net | CASIA-Net · Multi-PIE ±90° | accuracy: 65.6% to 77.1% | Helped |
| Add APRN to ResNet-101 | ResNet-101 · IJB-A FAR=10−3 | TAR: 93.1% to 95.1% | Helped |
| Add APRN to ResNet-152 | ResNet-152 · MegaFace Challenge 1 | rank1@106 accuracy, clean: 99.49% to 99.78% | Helped |
| Use stem features as the APRN discriminator input (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 96.17 | Hurt |
| Use stage1, stage2, stage3, or stage4 features rather than no discriminator input (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 96.33; 96.24 to 96.61; 96.24 to 96.61; 96.24 to 96.84 | Helped |
| Set loss weight α to 0.1 (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 96.61 | Helped |
| Set loss weight α to 0.2 (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 96.84 | Helped |
| Set loss weight α to 0.3 (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 96.31 | Helped |
| Set loss weight α to 0.4 (ablation) | APRN with ResNet-101 · CFP-FP | accuracy: 96.24 to 95.77 | Hurt |
| Use both parts of the adversarial regression loss instead of only the landmark-estimation or frontal-mean-face objective (ablation) | APRN with ResNet-101 · IJB-A FAR=10−4 | TAR: 88.30 to 90.10; 87.81 to 90.10 | Helped |
Closest papers by meaning
- Towards Pose Invariant Face Recognition in the Wild (CVPR 2018 · similarity 0.91)
- Disentangled Representation Learning GAN for Pose-Invariant Face Recognition (CVPR 2017 · similarity 0.88)
- A 3D GAN for Improved Large-Pose Facial Recognition (CVPR 2021 · similarity 0.87)
- Pose-Aware Face Recognition in the Wild (CVPR 2016 · similarity 0.86)
- Neural Refinement for Absolute Pose Regression with Feature Synthesis (CVPR 2024 · similarity 0.85)
- Towards Robust 3D Pose Transfer with Adversarial Learning (CVPR 2024 · 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.