Research map / Learning theory and algorithms / Matrix and tensor methods
Intrinsic Grassmann Averages for Online Linear, Robust and Nonlinear Subspace Learning
Rudrasis Chakraborty, Liu Yang, Søren Hauberg, Baba C. Vemuri
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 RIGA instead of Oja's rule for online PCA on body shape meshes | online linear subspace estimation · Human body shape meshes | average reconstruction error (mm): 16.8 mm -> 1.0 mm | Improved |
| use RIGA instead of EM PCA for online PCA on body shape meshes | online linear subspace estimation · Human body shape meshes | average reconstruction error (mm): 1.9 mm -> 1.0 mm | Improved |
| use RIGA instead of EM PCA on movie frames | online linear subspace estimation · Santa Claus Conquers the Martians frames | average reconstruction error: 0.025 -> 0.023 | No clear effect |
| use RIGA instead of Oja's rule on movie frames | online linear subspace estimation · Santa Claus Conquers the Martians frames | average reconstruction error: 0.054 -> 0.023 | Improved |
| use online KRIGA instead of KPCA with eigs for kernel subspace estimation | kernel subspace estimation · synthetic data, K=5 | computation time and ARE | Mixed |
| reduce the approximation dimension M of online KRIGA from 500 to 50 (ablation) | online KRIGA · synthetic concentric circles and 3D data | cluster separation: M=500 almost as good as KPCA; M=50 poor | Hurt |
| use RRIGA instead of GRASTA and DHR-PCA for robust PCA | robust subspace estimation · UCSD anomaly detection, reconstruction error and time | reconstruction error | Improved |
| use RRIGA instead of DHR-PCA and GRASTA at larger K | robust subspace estimation · Extended YaleB faces | reconstruction error | Mixed |
| use robust PCA (RRIGA) instead of standard PCA for background separation | Wallflower background separation · Camouflage and Waving Tree, 5% and 10% outliers | background separation quality | Improved |
Closest papers by meaning
- Intrinsic Grassmann Averages for Online Linear and Robust Subspace Learning (CVPR 2017 · similarity 0.97)
- Fun with Flags: Robust Principal Directions via Flag Manifolds (CVPR 2024 · similarity 0.89)
- Random features for Grassmannian kernel approximation with bounded rank-one projections (NeurIPS 2026 · similarity 0.88)
- A Geometric Analysis of PCA (NeurIPS 2025 · similarity 0.86)
- Principle Component Trees and Their Persistent Homology (AAAI 2024 · similarity 0.86)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.86)
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.
