Research map / Learning theory and algorithms / Matrix and tensor methods
High Order Tensor Formulation for Convolutional Sparse Coding
Adel Bibi, Bernard Ghanem
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 |
|---|---|---|---|
| increase number of filters K from 20 to 100 (ablation) | TCSC · Colored image reconstruction, 29 test images (TCSC, lambda=20, ~64% test sparsity) | PSNR (dB): 70.63 -> 104.70 | Improved |
| increase K from 80 to 100 (ablation) | TCSC · Colored image reconstruction, 29 test images (TCSC, lambda=20, ~64% test sparsity) | PSNR (dB): 104.43 -> 104.70 | No clear effect |
| increase SCSC training sparsity by retraining with lambda=5 instead of 1 (ablation) | SCSC · Colored image reconstruction, SCSC vs TCSC | training sparsity (%): 97.77 -> 98.53 | Mixed |
| train TCSC and SCSC at the same high training sparsity (lambda=20 for both) (ablation) | TCSC vs SCSC · Colored image reconstruction, 29 test images | PSNR vs testing sparsity: 99.55% vs 99.22% training sparsity | Improved |
| use TCSC dictionary with single-channel coding (TDSC) instead of SCSC dictionary (ablation) | SCSC coding · Colored image reconstruction, 29 test images | PSNR | Improved |
| replace independent per-channel CSC with tensor CSC | SCSC · Colored image reconstruction (matched training sparsity) | parameter count: n1 = 3 times fewer parameters | Improved |
Closest papers by meaning
- Consensus Convolutional Sparse Coding (ICCV 2017 · similarity 0.91)
- Convolutional Dictionary Learning via Local Processing (ICCV 2017 · similarity 0.89)
- Online Convolutional Sparse Coding with Sample-Dependent Dictionary (ICML 2018 · similarity 0.89)
- Distribution-Aware Tensor Decomposition for Compression of Convolutional Neural Networks (NeurIPS 2025 · similarity 0.87)
- Deep Rank-One Tensor Functional Factorization for Multi-Dimensional Data Recovery (AAAI 2025 · similarity 0.86)
- Coarse-To-Fine Tensor Trains for Compact Visual Representations (ICML 2024 · 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.
