Research map / Label-efficient and robust learning / Recognition and long-tail
Side Information for Face Completion: A Robust PCA Approach
Niannan Xue, Jiankang Deng, Shiyang Cheng, Yannis Panagakis, Stefanos P. Zafeiriou
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 side information to PCP (ablation) | PCP · 4DFAB subject #2 | PSNR (dB): 26.75 -> 30.63 | Improved |
| add side information and features to PCP (ablation) | PCP · 4DFAB subject #1 | PSNR (dB): 35.99 -> 41.48 | Improved |
| add features (subspaces) on top of PCPSM (ablation) | PCPSM · 4DFAB subject #5 | PSNR (dB): 32.80 -> 34.80 | Improved |
| add side information to PCP (ablation) | PCP · VGG verification | accuracy (%): 97.15 -> 98.78 | Improved |
| add features and side information from GAN to PCP (ablation) | PCP · CFP identification | rank-1 accuracy (%): 94.60 -> 97.60 | Improved |
| add side information on top of features (PCPSFM over LRR) (ablation) | LRR · VGG identification | rank-1 accuracy (%): 97.55 -> 98.10 | Improved |
| add GAN-derived side information to PCP (ablation) | PCP · YTF | verification accuracy (%): 95.25 -> 95.94 | Improved |
| fuse PCPSFM with GAN output (ablation) | GAN · YTF | verification accuracy (%): 95.68 -> 96.58 | Improved |
| add side information (PCPSFM) over feature-only PCPF/LRR (ablation) | PCPF · synthetic phase transition (full observation) | recovery region | Improved |
| add side information when 10% of elements are missing (ablation) | PCP · synthetic phase transition (partial observation) | recovery region | Improved |
| use imperfect side information (average of 10 other subjects) (ablation) | PCPSM/PCPSFM · Extended Yale B denoising | visual denoising | Improved |
| use GAN alone instead of PCPSM or PCP | PCP · 300VW qualitative UV completion | visual quality | Mixed |
Closest papers by meaning
- Side Information in Robust Principal Component Analysis: Algorithms and Applications (ICCV 2017 · similarity 0.90)
- Robust Principal Component Analysis with Side Information (ICML 2016 · similarity 0.87)
- UV-GAN: Adversarial Facial UV Map Completion for Pose-Invariant Face Recognition (CVPR 2018 · similarity 0.86)
- Partial Sum Minimization of Singular Values in Robust PCA: Algorithm and Applications (TPAMI 2016 · similarity 0.86)
- Additive Component Analysis (CVPR 2017 · similarity 0.85)
- A Nonconvex Projection Method for Robust PCA (AAAI 2019 · 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.
