Research map / Learning theory and algorithms / Clustering
Highly-Economized Multi-View Binary Compression for Scalable Image Clustering
Zheng Zhang, Li Liu, Jie Qin, Fan Zhu, Fumin Shen, Yong Xu, Ling Shao, Heng Tao Shen
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 joint binary coding and clustering instead of two separate steps (HSIC vs HSIC-TS) (ablation) | HSIC · ImageNet-10 MulView | ACC: 0.3568 -> 0.3865 | Improved |
| use l21-norm instead of F-norm for binary clustering (HSIC vs HSIC-F) (ablation) | HSIC · ImageNet-10 MulView | ACC: 0.3749 -> 0.3865 | Improved |
| joint learning vs two separate steps (ablation) | HSIC · Cifar-10 MulView | ACC: 0.3742 -> 0.3951 | Improved |
| joint learning vs two separate steps (ablation) | HSIC · YouTube Faces MulView | ACC: 0.5974 -> 0.6547 | Improved |
| l21-norm vs F-norm (ablation) | HSIC · NUS-WIDE MulView | ACC: 0.1639 -> 0.1661 | Improved |
| use multi-view features instead of single view LBP (ablation) | HSIC · ImageNet-10 | ACC: 0.2275 -> 0.3865 | Improved |
| binary codes instead of real-valued k-means features | HSIC · ImageNet-10 time | time (s): 201 -> 5 | Improved |
| binary codes instead of real-valued k-means features | HSIC · ImageNet-10 memory | memory: 361 MB -> 2.73 MB | Improved |
| binary clustering on NUS-WIDE with 5 views instead of k-means | HSIC · NUS-WIDE | time and memory: 961 MB -> 5.52 MB; 81 seconds | Improved |
| use HSIC with fast Hamming distance instead of ITQ+bk-means on YouTube Faces MulView | HSIC · YouTube Faces | time (s): 110 -> 139 | Hurt |
| remove balanced and independence constraints (HSIC-U) (ablation) | HSIC · Cifar-10 component analysis | clustering performance | Improved |
| increase code length beyond 32 bits | HSIC · Cifar-10 | clustering performance: 32 bits | Improved |
Closest papers by meaning
- Binary Multi-View Clustering (TPAMI 2018 · similarity 0.91)
- Clustering by Maximizing Mutual Information Across Views (ICCV 2021 · similarity 0.86)
- Towards Resource-friendly, Extensible and Stable Incomplete Multi-view Clustering (ICML 2024 · similarity 0.85)
- Another Way to the Top: Exploit Contextual Clustering in Learned Image Coding (AAAI 2024 · similarity 0.85)
- Bifurcate then Alienate: Incomplete Multi-view Clustering via Coupled Distribution Learning with Linear Overhead (ICML 2025 · similarity 0.84)
- Medusa: A Multi-Scale High-order Contrastive Dual-Diffusion Approach for Multi-View Clustering (CVPR 2025 · similarity 0.84)
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.
