Research map / Label-efficient and robust learning / Recognition and long-tail
Weakly Supervised Deep Hyperspherical Quantization for Image Retrieval
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 WSDHQ instead of WDHT | weakly supervised image retrieval · MIR-FLICKR25K, 32 bits | MAP@5000: 0.724 -> 0.772 | Helped |
| use WSDHQ instead of WDHT | weakly supervised image retrieval · NUS-WIDE, 32 bits | MAP@5000: 0.692 -> 0.731 | Helped |
| add tag correlation graph, semantic enhancement, and synonym merging (ablation) | WSDHQ · MIR-FLICKR25K and NUS-WIDE | average MAP: +2.3% and +1.6% | Helped |
| apply L2 normalization to tag and image embeddings and learn on the hypersphere (ablation) | WSDHQ · MIR-FLICKR25K and NUS-WIDE | average MAP: +2.9% and +1.4% | Helped |
| use adaptive cosine margin loss instead of WDHT mini-batch hinge loss (ablation) | WSDHQ · MIR-FLICKR25K and NUS-WIDE | average MAP: +2.8% and +3.3% | Helped |
| jointly learn embeddings and quantization instead of using two stages (ablation) | WSDHQ · MIR-FLICKR25K and NUS-WIDE | average MAP: +3.2% and +3.5% | Helped |
Closest papers by meaning
- HiHPQ: Hierarchical Hyperbolic Product Quantization for Unsupervised Image Retrieval (AAAI 2024 · similarity 0.89)
- Weakly Supervised Deep Image Hashing Through Tag Embeddings (CVPR 2019 · similarity 0.88)
- Similarity Preserving Deep Asymmetric Quantization for Image Retrieval (AAAI 2019 · similarity 0.88)
- Supervised Quantization for Similarity Search (CVPR 2016 · similarity 0.87)
- Deep Visual-Semantic Quantization for Efficient Image Retrieval (CVPR 2017 · similarity 0.87)
- Learning Structured Representations with Hyperbolic Embeddings (NeurIPS 2024 · 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.