Research map / Label-efficient and robust learning / Semi-supervised learning
Fast Class-wise Updating for Online Hashing
Mingbao Lin, Rongrong Ji, Xiaoshuai Sun, Baochang Zhang, Feiyue Huang, Yonghong Tian, Dacheng Tao
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 class-wise updating (compared with using all classes each round) (ablation) | FCOH · CIFAR-10 (32-bit) | mAP: 0.691 -> 0.702 | Improved |
| add class-wise updating (compared with using all classes each round) (ablation) | FCOH · Places205 (32-bit) | mAP@1000: 0.255 -> 0.267 | Improved |
| add class-wise updating (compared with using all classes each round) (ablation) | FCOH · MNIST (32-bit) | mAP: 0.760 -> 0.786 | Improved |
| add semi-relaxation optimization instead of DCC (ablation) | FCOH · CIFAR-10 (32-bit) | mAP: 0.696 -> 0.702 | Improved |
| add semi-relaxation optimization instead of DCC (ablation) | FCOH · Places205 (32-bit) | mAP@1000: 0.262 -> 0.267 | Improved |
| add semi-relaxation optimization instead of DCC (ablation) | FCOH · MNIST (32-bit) | mAP: 0.777 -> 0.786 | Improved |
| add both class-wise updating and semi-relaxation to BSODH (ablation) | BSODH · MNIST (32-bit) | mAP: 0.747 -> 0.786 | Improved |
| use similar-item term lambda1 > 0 instead of 0 (ablation) | FCOH · CIFAR-10 (32-bit) | mAP: 0.565 -> 0.702 | Improved |
| use similar-item term lambda1 > 0 instead of 0 (ablation) | FCOH · Places205 (32-bit) | mAP@1000: 0.099 -> 0.267 | Improved |
| use dissimilar-item term lambda2 > 0 instead of 0 (ablation) | FCOH · CIFAR-10 (32-bit) | mAP: 0.554 -> 0.702 | Improved |
| use dissimilar-item term lambda2 > 0 instead of 0 (ablation) | FCOH · MNIST (32-bit) | mAP: 0.637 -> 0.786 | Improved |
| increase batch size n_t (ablation) | FCOH · CIFAR-10 and MNIST | mAP | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MIHash — MIHash: Online Hashing With Mutual Information (ICCV 2017) · better on 3 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Online Hashing with Efficient Updating of Binary Codes (AAAI 2020 · similarity 0.91)
- KALAHash: Knowledge-Anchored Low-Resource Adaptation for Deep Hashing (AAAI 2025 · similarity 0.86)
- Pairwise-Label-Based Deep Incremental Hashing with Simultaneous Code Expansion (AAAI 2024 · similarity 0.86)
- Semi-Supervised Online Cross-Modal Hashing (AAAI 2025 · similarity 0.86)
- Deep Graph Online Hashing for Multi-Label Image Retrieval (AAAI 2025 · similarity 0.85)
- Generalized Debiased Semi-Supervised Hashing for Large-Scale Image Retrieval (AAAI 2025 · similarity 0.85)
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.
