Research map / Image generation / Editing and style transfer
Neural Transformation Fields for Arbitrary-Styled Font Generation
Bin Fu, Junjun He, Jianjun Wang, Yu Qiao
Read the paperCheck 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 localized rather than universal style representation (ablation) | NTF · Unseen Fonts and Seen Contents | SSIM: 0.6102 -> 0.6299 | Improved |
| use localized rather than universal style representation (ablation) | NTF · Unseen Fonts and Unseen Contents | SSIM: 0.6331 -> 0.6533 | Improved |
| use localized rather than universal style representation (ablation) | NTF · Unseen Fonts and Seen Contents | FID: 30.46 -> 13.72 | Improved |
| use localized rather than universal style representation (ablation) | NTF · Unseen Fonts and Unseen Contents | FID: 33.12 -> 15.67 | Improved |
| add NTF-based transformation and rendering instead of the base generation process (ablation) | Font generation · Unseen Fonts and Seen Contents | SSIM: 0.6189 -> 0.6229 | Improved |
| use 10 rather than 5 sampling points (ablation) | NTF font rendering · Unseen Fonts and Seen Contents | SSIM: 0.6229 -> 0.6280 | Improved |
| use 15 rather than 10 sampling points (ablation) | NTF font rendering · Unseen Fonts and Seen Contents | SSIM: 0.6280 -> 0.6299 | Improved |
| increase sampling points from 15 to 20 (ablation) | NTF font rendering · Unseen Fonts and Seen Contents | SSIM: 0.6299 -> 0.6260 | Hurt |
| add NTF-based transformation and rendering instead of the base generation process (ablation) | Font generation · Unseen Fonts and Unseen Contents | SSIM: 0.6413 -> 0.6466 | Improved |
| use 10 rather than 5 sampling points (ablation) | NTF font rendering · Unseen Fonts and Unseen Contents | SSIM: 0.6466 -> 0.6503 | Improved |
| use 15 rather than 10 sampling points (ablation) | NTF font rendering · Unseen Fonts and Unseen Contents | SSIM: 0.6503 -> 0.6533 | Improved |
| increase sampling points from 15 to 20 (ablation) | NTF font rendering · Unseen Fonts and Unseen Contents | SSIM: 0.6533 -> 0.6473 | Hurt |
Closest papers by meaning
- Generate Like Experts: Multi-Stage Font Generation by Incorporating Font Transfer Process into Diffusion Models (CVPR 2024 · similarity 0.93)
- FontAnimate: High Quality Few-shot Font Generation via Animating Font Transfer Process (ICCV 2025 · similarity 0.91)
- IF-Font: Ideographic Description Sequence-Following Font Generation (NeurIPS 2024 · similarity 0.89)
- FontDiffuser: One-Shot Font Generation via Denoising Diffusion with Multi-Scale Content Aggregation and Style Contrastive Learning (AAAI 2024 · similarity 0.87)
- VQ-FONT: Few-Shot Font Generation with Structure-Aware Enhancement and Quantization (AAAI 2024 · similarity 0.86)
- DeepCalliFont: Few-Shot Chinese Calligraphy Font Synthesis by Integrating Dual-Modality Generative Models (AAAI 2024 · similarity 0.86)
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.
