Research map / Graph learning / Graph algorithms and clustering
Deep Supervised Hashing With Anchor Graph
Yudong Chen, Zhihui Lai, Yujuan Ding, Kaiyi Lin, Wai Keung Wong
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 |
|---|---|---|---|
| obtain training codes directly from Algorithm 1 instead of the network output (DAGH-2 vs DAGH-1) (ablation) | DAGH · CIFAR-10, 12 bits | MAP: 0.881 -> 0.934 | Improved |
| obtain training codes directly from Algorithm 1 instead of the network output (DAGH-2 vs DAGH-1) (ablation) | DAGH · Fashion-MNIST, 12 bits | MAP: 0.870 -> 0.932 | Improved |
| obtain training codes directly from Algorithm 1 instead of the network output (DAGH-2 vs DAGH-1) (ablation) | DAGH · NUS-WIDE, 12 bits | MAP: 0.707 -> 0.760 | Improved |
| add the deep hashing term R2 to the regression term R1 (ablation) | DAGH · NUS-WIDE, 12 bits | MAP: 0.739 -> 0.772 | Improved |
| add the deep hashing term R2 to the regression term R1 (ablation) | DAGH · NUS-WIDE, 48 bits | MAP: 0.762 -> 0.787 | Improved |
| add the discriminative code term R3 to the regression term R1 (ablation) | DAGH · NUS-WIDE, 12 bits | MAP: 0.739 -> 0.740 | No clear effect |
| add the discriminative code term R3 to the regression term R1 (ablation) | DAGH · NUS-WIDE, 48 bits | MAP: 0.762 -> 0.790 | Improved |
| add R3 on top of R1+R2 (ablation) | DAGH · NUS-WIDE, 12 bits | MAP: 0.772 -> 0.760 | Hurt |
| add R3 on top of R1+R2 (ablation) | DAGH · NUS-WIDE, 48 bits | MAP: 0.787 -> 0.802 | Improved |
| train DPSH on all training samples instead of a 5,000-image subset | DPSH · CIFAR-10, 12 bits | MAP: 0.661 -> 0.925 (time 58.5m -> 842.3m) | Improved |
| train DSDH on all training samples instead of a subset | DSDH · CIFAR-10, 12 bits | MAP: 0.735 -> 0.931 (time 58.9m -> 846.1m) | Improved |
| vary the number of anchors (ablation) | DAGH · CIFAR-10 | MAP | No clear effect |
Closest papers by meaning
- Deep Supervised Discrete Hashing (NeurIPS 2017 · similarity 0.91)
- On Trivial Solution and High Correlation Problems in Deep Supervised Hashing (AAAI 2018 · similarity 0.88)
- Learning to Hash with Binary Deep Neural Network (ECCV 2016 · similarity 0.87)
- Deep Hashing with Active Pairwise Supervision (ECCV 2020 · similarity 0.87)
- Asymmetric Discrete Graph Hashing (AAAI 2017 · similarity 0.87)
- Deep Graph Online Hashing for Multi-Label Image Retrieval (AAAI 2025 · similarity 0.87)
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.
