Research map / Architectures and efficiency / RNNs and normalization
Deep Unsupervised Image Hashing by Maximizing Bit Entropy
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 |
|---|---|---|---|
| Replace sign-based hashing with the Bi-Half layer | Unsupervised deep hashing · Cifar-10(I) | mAP@1000: 44.80 -> 56.10; 47.20 -> 57.60; 50.10 -> 59.50 | Helped |
| Replace sign-based hashing with the Bi-Half layer | Unsupervised deep hashing · Nus-wide(I) | mAP@All: 55.49 -> 65.12; 57.47 -> 66.31; 60.93 -> 67.26 | Helped |
| Replace the strongest listed baseline with the Bi-Half layer | Unsupervised deep hashing · Flickr25k | mAP@All: 69.64 -> 71.42; 70.56 -> 72.35; 70.75 -> 73.10 | Helped |
| Use the Bi-Half layer instead of DistillHash | Unsupervised deep hashing · Nus-wide(II) | mAP@All: 66.67 -> 67.12; 67.52 -> 68.05; 67.69 -> 68.21 | Mixed |
| Replace sign-based hashing with the Bi-Half layer | Unsupervised deep hashing · Cifar-10(II) | mAP@All: 28.71 -> 42.87; 31.72 -> 43.29; 35.47 -> 44.13 | Helped |
| Replace sign-based hashing with the Bi-Half layer | ResNet-34 video hashing · Ucf-101 | mAP@100: 45.49 -> 50.83; 57.24 -> 60.30; 64.77 -> 65.89 | Helped |
| Replace sign-based hashing with the Bi-Half layer | ResNet-34 video hashing · Hmdb-51 | mAP@100: 30.32 -> 34.21; 37.55 -> 38.67 | Helped |
| Add entropy regularization instead of using the Bi-Half layer (ablation) | Mnist autoencoder hashing · Mnist | mAP | Hurt |
| Use the sign layer instead of the Bi-Half layer (ablation) | Mnist autoencoder hashing · Mnist | reconstruction loss | Hurt |
| Replace the Bi-Half layer with median-based translation (ablation) | Flickr25k hashing · Flickr25k | mAP | Hurt |
| Use smaller mini-batches for median-based translation (ablation) | Median-based hashing · Flickr25k | mAP | Hurt |
| Replace Bi-Half coding with sign coding (ablation) | Cifar-10(I) hashing · Cifar-10(I) | mAP | Hurt |
Closest papers by meaning
- Unsupervised Deep Generative Adversarial Hashing Network (CVPR 2018 · similarity 0.89)
- Deep Hashing Network for Efficient Similarity Retrieval (AAAI 2016 · similarity 0.89)
- Deep Unsupervised Hashing via External Guidance (ICML 2025 · similarity 0.88)
- An Asymmetric Augmented Self-Supervised Learning Method for Unsupervised Fine-Grained Image Hashing (CVPR 2024 · similarity 0.86)
- Pairwise-Label-Based Deep Incremental Hashing with Simultaneous Code Expansion (AAAI 2024 · similarity 0.86)
- Learning Attribute-Aware Hash Codes for Fine-Grained Image Retrieval via Query Optimization (ICML 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.