Research map / Label-efficient and robust learning / Recognition and long-tail
Cross Euclidean-to-Riemannian Metric Learning with Application to Face Recognition from Video
Zhiwu Huang, Ruiping Wang, Shiguang Shan, Luc Van Gool, Xilin Chen
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 |
|---|---|---|---|
| use RBF kernel instead of linear kernel (ablation) | CERML-ES · COX V2S V1-S (deep) | accuracy (%): 84.67 -> 86.21 | Improved |
| use RBF kernel instead of linear kernel (ablation) | CERML-EG · COX V2S V1-S (deep) | accuracy (%): 85.36 -> 85.71 | No clear effect |
| use RBF kernel instead of linear kernel (ablation) | CERML-EA · COX V2S V1-S (deep) | accuracy (%): 86.17 -> 86.40 | No clear effect |
| use RBF kernel instead of linear kernel (ablation) | CERML-ES · COX V2V V3-V2 (deep) | accuracy (%): 98.77 -> 97.46 | Hurt |
| use RBF kernel instead of linear kernel (ablation) | CERML-EG · COX V2V V2-V1 (deep) | accuracy (%): 97.84 -> 98.77 | Improved |
| use CERML-ES instead of Euclidean metric learning LMNN | CERML-ES · COX V2S V1-S (gray) | accuracy (%): 34.44 -> 51.41 | Improved |
| use CERML-ES instead of kernel cross-metric KGMA-ES | CERML-ES · COX V2S V1-S (gray) | accuracy (%): 41.89 -> 51.41 | Improved |
| use CERML-EG instead of CDL | CERML-EG · COX V2V V2-V1 (deep) | accuracy (%): 95.53 -> 97.77 | Improved |
| use CERML-ES instead of VGGDeepFace on PaSC control V2V | CERML-ES · PaSC V2V control (deep) | accuracy (%): 78.82 -> 79.92 | Improved |
| use CERML-EG instead of PML | CERML-EG · YTC (deep) | accuracy (%): 86.98 -> 87.62 | Improved |
| compare CERML-ES with FaceNet | CERML-ES · YTF (deep) | accuracy (%): 95.1 vs 93.44 | Hurt |
| use CERML-EG instead of LFDA on PaSC V2S control | CERML-EG · PaSC V2S control (deep) | accuracy (%): 72.44 -> 63.34 | Hurt |
Closest papers by meaning
- Approximate Log-Hilbert-Schmidt Distances Between Covariance Operators for Image Classification (CVPR 2016 · similarity 0.87)
- Dimensionality Reduction on SPD Manifolds: The Emergence of Geometry-Aware Methods (TPAMI 2017 · similarity 0.87)
- A Novel Space-Time Representation on the Positive Semidefinite Cone for Facial Expression Recognition (ICCV 2017 · similarity 0.87)
- Task-Induced Riemannian Metrics for Vision Transformer Feature Spaces (NeurIPS 2026 · similarity 0.85)
- Intrinsic Riemannian Cross-covariance for Manifold-valued Random Objects (NeurIPS 2026 · similarity 0.84)
- Euclidean Embedding of Data Using Local Distances (NeurIPS 2026 · 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.
