Research map / Adversarial robustness and security / Attacks on text and graphs
Prototype-Supervised Adversarial Network for Targeted Attack of Deep Hashing
Xunguang Wang, Zheng Zhang, Baoyuan Wu, Fumin Shen, Guangming Lu
Check 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 |
|---|---|---|---|
| Use ProS-GAN instead of DHTA | DPSH hashing retrieval · FLICKR-25K, 32 bits | t-MAP: 88.35 -> 91.10 | Improved |
| Use ProS-GAN instead of DHTA | DPSH hashing retrieval · NUS-WIDE, 32 bits | t-MAP: 75.65 -> 78.25 | Improved |
| Use ProS-GAN instead of DHTA | DPSH hashing retrieval · MS-COCO, 32 bits | t-MAP: 63.22 -> 71.65 | Improved |
| Add prototype-code supervision using the Hamming distance loss (ablation) | ProS-GAN · NUS-WIDE | t-MAP | Improved |
| Add the discriminator classification module (ablation) | ProS-GAN · NUS-WIDE, varying hash bits | t-MAP | Improved |
| Increase the reconstruction-loss weight α (ablation) | ProS-GAN · NUS-WIDE, 32 bits | perceptibility | Improved |
| Increase the reconstruction-loss weight α (ablation) | ProS-GAN · NUS-WIDE, 32 bits | t-MAP | Hurt |
| Use learned prototype codes instead of anchor codes | targeted hashing attack · FLICKR-25K, 32 bits | t-MAP: 88.68 -> 92.91 | Improved |
| Use learned prototype codes instead of anchor codes | targeted hashing attack · NUS-WIDE, 32 bits | t-MAP: 78.40 -> 81.34 | Improved |
| Use learned prototype codes instead of anchor codes | targeted hashing attack · MS-COCO, 32 bits | t-MAP: 64.41 -> 71.93 | Improved |
Closest papers by meaning
- Semi-Supervised Generative Adversarial Hashing for Image Retrieval (ECCV 2018 · similarity 0.90)
- Targeted Attack for Deep Hashing based Retrieval (ECCV 2020 · similarity 0.90)
- SSAH: Semi-Supervised Adversarial Deep Hashing with Self-Paced Hard Sample Generation (AAAI 2020 · similarity 0.88)
- Deep Supervised Adversarial Robust Hashing for Retrieval (TPAMI 2026 · similarity 0.88)
- HUANG: A Robust Diffusion Model-based Targeted Adversarial Attack Against Deep Hashing Retrieval (AAAI 2025 · similarity 0.87)
- Once and for All: Universal Transferable Adversarial Perturbation against Deep Hashing-Based Facial Image Retrieval (AAAI 2024 · 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.
