Research map / Image generation / Autoregressive and latent generation
Inverse Discriminative Networks for Handwritten Signature Verification
Ping Wei, Huan Li, Ping Hu
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 |
|---|---|---|---|
| add inverse streams and multi-path attention modules (ablation) | Double Stream signature verifier · CSD test | Acc: 88.26 -> 90.17 | Improved |
| use one concatenated-image stream instead of the four-stream IDN (ablation) | IDN · CSD test | Acc: 90.17 -> 88.06 | Hurt |
| add inverse streams and multi-path attention modules (ablation) | Double Stream signature verifier · CEDAR test | EER: 4.86 -> 3.62 | Improved |
| use one concatenated-image stream instead of the four-stream IDN (ablation) | IDN · CEDAR test | EER: 3.62 -> 10.0 | Hurt |
| add inverse streams and multi-path attention modules (ablation) | Double Stream signature verifier · BHSig-B test | Acc: 91.14 -> 95.32 | Improved |
| use one concatenated-image stream instead of the four-stream IDN (ablation) | IDN · BHSig-B test | Acc: 95.32 -> 88.76 | Hurt |
| add inverse streams and multi-path attention modules (ablation) | Double Stream signature verifier · BHSig-H test | Acc: 90.62 -> 93.04 | Improved |
| use one concatenated-image stream instead of the four-stream IDN (ablation) | IDN · BHSig-H test | Acc: 93.04 -> 87.44 | Hurt |
| train on one language and test on another | IDN · cross-language signature verification | accuracy: 90.17 -> 50.0 | Hurt |
| train on Bengali signatures and test on Hindi signatures | IDN · BHSig-B to BHSig-H | accuracy: 95.32 -> 74.30 | Hurt |
Closest papers by meaning
- SynSig2Vec: Forgery-Free Learning of Dynamic Signature Representations by Sigma Lognormal-Based Synthesis and 1D CNN (TPAMI 2021 · similarity 0.85)
- Deep Signature Transforms (NeurIPS 2019 · similarity 0.84)
- Static-Dynamic Interaction Networks for Offline Signature Verification (AAAI 2021 · similarity 0.84)
- ZeroMark: Towards Dataset Ownership Verification without Disclosing Watermark (NeurIPS 2024 · similarity 0.80)
- Probabilistic Signature Inversion: Learning Conditional Distributions from Truncated Signatures (NeurIPS 2026 · similarity 0.80)
- Beyond Slow Signs in High-fidelity Model Extraction (NeurIPS 2024 · similarity 0.80)
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.
