Research map / Learning theory and algorithms / Prediction and decision losses
Deep Hashing With Minimal-Distance-Separated Hash Centers
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 optimized hash centers with a constrained minimum Hamming distance instead of OrthoHash | Deep hashing with minimal-distance-separated centers · ImageNet | MAP@1000: 0.8540 -> 0.8639; 0.8792 -> 0.8863; 0.8936 -> 0.9019 | Helped |
| Use optimized hash centers with a constrained minimum Hamming distance instead of OrthoHash | Deep hashing with minimal-distance-separated centers · NABirds | MAP@All: 0.6366 -> 0.6977; 0.7243 -> 0.7417; 0.7544 -> 0.7619 | Helped |
| Use optimized hash centers with a constrained minimum Hamming distance instead of OrthoHash | Deep hashing with minimal-distance-separated centers · Stanford Cars | MAP@All: 0.8012 -> 0.8579; 0.8490 -> 0.8731; 0.8676 -> 0.8814 | Helped |
| Add optimized, minimum-distance-separated centers instead of Hadamard and Bernoulli-sampled centers (ablation) | Deep hashing network · ImageNet, Stanford Cars, and NABirds | MAP: ImageNet: 0.8535 -> 0.8639; Stanford Cars: 0.7977 -> 0.8579; NABirds: 0.6544 -> 0.6977 | Helped |
| Add the similar-pair loss LP (ablation) | Proposed deep hashing network · ImageNet, Stanford Cars, and NABirds | MAP: ImageNet: 0.8561 -> 0.8639; Stanford Cars: 0.8221 -> 0.8579; NABirds: 0.6810 -> 0.6977 | Helped |
| Add the quantization loss LQ (ablation) | Proposed deep hashing network · ImageNet, Stanford Cars, and NABirds | MAP: ImageNet: 0.8601 -> 0.8639; Stanford Cars: 0.8451 -> 0.8579; NABirds: 0.6937 -> 0.6977 | Helped |
| Add both LP and LQ (ablation) | Proposed deep hashing network · ImageNet, Stanford Cars, and NABirds | MAP: ImageNet: 0.8558 -> 0.8639; Stanford Cars: 0.8182 -> 0.8579; NABirds: 0.6715 -> 0.6977 | Helped |
| Replace the CSQ center loss with the proposed center loss LC (ablation) | Deep hashing network with the same hash centers · ImageNet, Stanford Cars, and NABirds | MAP: ImageNet: 0.8457 -> 0.8639; Stanford Cars: 0.7888 -> 0.8579; NABirds: 0.6351 -> 0.6977 | Helped |
Closest papers by meaning
- Central Similarity Quantization for Efficient Image and Video Retrieval (CVPR 2020 · similarity 0.94)
- Deep Hashing Network for Efficient Similarity Retrieval (AAAI 2016 · similarity 0.91)
- One Loss for Quantization: Deep Hashing With Discrete Wasserstein Distributional Matching (CVPR 2022 · similarity 0.89)
- Deep Supervised Hashing for Fast Image Retrieval (CVPR 2016 · similarity 0.89)
- One Loss for All: Deep Hashing with a Single Cosine Similarity based Learning Objective (NeurIPS 2021 · similarity 0.89)
- Pairwise-Label-Based Deep Incremental Hashing with Simultaneous Code Expansion (AAAI 2024 · similarity 0.88)
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.