Research map / Image generation / GANs
StrokeGAN: Reducing Mode Collapse in Chinese Font Generation via Stroke Encoding
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 |
|---|---|---|---|
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Regular Script → Shu | content accuracy (%): 89.56 -> 90.48 | Helped |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Regular Script → Shu | recognition accuracy (%): 89.36 -> 90.52 | Helped |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Regular Script → Shu | stroke error (× 10⁻²): 6.79 -> 5.69 | Helped |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Regular Script → Huawen Amber | stroke error (× 10⁻²): 8.71 -> 7.20 | Helped |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Regular Script → Imitated Song | stroke error (× 10⁻²): 7.33 -> 5.55 | Helped |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Imitated Song → Black | content accuracy (%): 89.56 -> 88.96 | Hurt |
| Add one-bit stroke encoding and stroke-encoding reconstruction loss to CycleGAN | CycleGAN · Imitated Song → Black | stroke error (× 10⁻²): 8.06 -> 6.72 | Helped |
| Use StrokeGAN rather than CycleGAN | Chinese font generation · All nine reported font translations | stroke error (× 10⁻²) | Helped |
| Use one-bit stroke encoding for characters with shared encodings | StrokeGAN · Similar Chinese characters with the same stroke encodings | character distinguishability | Helped |
Closest papers by meaning
- FontRL: Chinese Font Synthesis via Deep Reinforcement Learning (AAAI 2021 · similarity 0.89)
- VQ-FONT: Few-Shot Font Generation with Structure-Aware Enhancement and Quantization (AAAI 2024 · similarity 0.88)
- Decoupling Layout from Glyph in Online Chinese Handwriting Generation (ICLR 2025 · similarity 0.88)
- DeepCalliFont: Few-Shot Chinese Calligraphy Font Synthesis by Integrating Dual-Modality Generative Models (AAAI 2024 · similarity 0.87)
- Generate Like Experts: Multi-Stage Font Generation by Incorporating Font Transfer Process into Diffusion Models (CVPR 2024 · similarity 0.86)
- FontDiffuser: One-Shot Font Generation via Denoising Diffusion with Multi-Scale Content Aggregation and Style Contrastive Learning (AAAI 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.