Research map / Learning theory and algorithms / Matrix and tensor methods
Low-Rank Tensor Completion by Approximating the Tensor Average Rank
Zhanliang Wang, Junyu Dong, Xinguo Liu, Xueying Zeng
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 |
|---|---|---|---|
| replace TNN-DCT with the nonconvex tensor average rank surrogate | DCT-based low-rank tensor completion · Standard images, sampling rate 0.2 | PSNR: 22.17 -> 23.01 | Improved |
| replace TNN-DCT with the nonconvex tensor average rank surrogate | DCT-based low-rank tensor completion · CBSD68, sampling rate 0.8 | PSNR: 44.51 -> 44.74 | Improved |
| replace TNN-DCT with the nonconvex surrogate on a dynamic video | DCT-based low-rank tensor completion · Container video, sampling rate 0.4 | PSNR: 44.35 -> 45.66 | Improved |
| replace TNN-DCT with the nonconvex surrogate | DCT-based low-rank tensor completion · Akiyo video, sampling rate 0.2 | PSNR: 36.40 -> 37.39 | Improved |
| use the closed-form hard thresholding algorithm instead of TNN-DCT's solver | tensor completion · Standard images | runtime (s): 4.59 -> 0.91 | Improved |
| replace TNN-DCT with the nonconvex surrogate on Container | tensor completion · Container video | runtime (s): 27.02 -> 3.86 | Improved |
| use nonconvex surrogate instead of TNN-DCT in synthetic recovery at sampling rate 0.5 | synthetic tensor completion (n=40 and n=50) · Synthetic recovery phase transition | max recoverable rank: four higher than that of TNN-DCT | Improved |
| add coarse-to-fine continuation (adaptively decreasing lambda) instead of fixed small lambda (ablation) | proposed tensor completion algorithm · Synthetic 10x10x10 tensor, average rank 1, sampling rate 0.5 | relative square error and convergence speed | Improved |
| SPC at low sampling rate versus other baselines | SPC · Standard images, sampling rate 0.2 | PSNR: 24.51 vs 22.17 (TNN-DCT) | Mixed |
Closest papers by meaning
- Low-Rank Tensor Completion With a New Tensor Nuclear Norm Induced by Invertible Linear Transforms (CVPR 2019 · similarity 0.94)
- Nonconvex Low-Rank Tensor Completion from Noisy Data (NeurIPS 2019 · similarity 0.94)
- Low-Rank Tensor Completion: A Pseudo-Bayesian Learning Approach (ICCV 2017 · similarity 0.92)
- Efficient Nonconvex Regularized Tensor Completion with Structure-aware Proximal Iterations (ICML 2019 · similarity 0.92)
- Fast Tensor Completion via Approximate Richardson Iteration (ICML 2025 · similarity 0.91)
- Low-Rank Tensor Learning by Generalized Nonconvex Regularization (TPAMI 2026 · similarity 0.91)
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.
