Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Subspace Clustering with Data Augmentation
Mahdi Abavisani, Alireza Naghizadeh, Dimitris Metaxas, Vishal Patel
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 data augmentation only (no consistency loss) (ablation) | DSC · Extended Yale B | clustering error (%): 2.67 -> 3.10 | Hurt |
| add data augmentation only (no consistency loss) (ablation) | MLRDSC · Extended Yale B | clustering error (%): 1.36 -> 2.84 | Hurt |
| add consistency loss without augmentation (ablation) | DSC · Extended Yale B | clustering error (%): 2.67 -> 2.56 | Improved |
| add consistency loss without augmentation (ablation) | MLRDSC · Extended Yale B | clustering error (%): 1.36 -> 0.95 | Improved |
| add augmentation together with consistency loss (ablation) | DSC · Extended Yale B | clustering error (%): 2.56 -> 1.92 | Improved |
| add augmentation together with consistency loss (ablation) | MLRDSC · Extended Yale B | clustering error (%): 0.95 -> 0.82 | Improved |
| use MLRDSC backbone instead of DSC (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 1.92 -> 0.82 | Improved |
| use searched policies instead of Random LR Flips (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 1.32 -> 0.82 | Improved |
| use searched policies instead of Cut-out (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 2.88 -> 0.82 | Improved |
| use searched policies instead of common practitioner augmentations (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 2.96 -> 0.82 | Improved |
| use searched policies instead of AutoAugment ImageNet policy (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 5.96 -> 0.82 | Improved |
| use searched policies instead of AutoAugment SVHN policy (ablation) | MLRDSC-DA · Extended Yale B | clustering error (%): 11.31 -> 0.82 | Improved |
Closest papers by meaning
- Exploring a Principled Framework for Deep Subspace Clustering (ICLR 2025 · similarity 0.91)
- LRSC: Learning Representations for Subspace Clustering (AAAI 2021 · similarity 0.91)
- Deep Subspace Clustering Networks (NeurIPS 2017 · similarity 0.90)
- Neural Collaborative Subspace Clustering (ICML 2019 · similarity 0.89)
- Efficient Deep Embedded Subspace Clustering (CVPR 2022 · similarity 0.89)
- A Critique of Self-Expressive Deep Subspace Clustering (ICLR 2021 · similarity 0.89)
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.
