Research map / Video / Video generation
FontAnimate: High Quality Few-shot Font Generation via Animating Font Transfer Process
Bin Fu, Zixuan Wang, Kainan Yan, Shitian Zhao, Qi Qin, Jie Wen, Junjun He, Peng Gao
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 |
|---|---|---|---|
| add temporal part and frame-index (I2V) to I2I font model (ablation) | FontAnimate I2I · UFUC (unseen fonts, unseen contents) | RMSE: 0.2575 -> 0.2483 | Improved |
| add temporal part and frame-index (I2V) to I2I font model (ablation) | FontAnimate I2I · UFUC (unseen fonts, unseen contents) | LPIPS: 0.1448 -> 0.1340 | Improved |
| add temporal part and frame-index (I2V) to I2I font model (ablation) | FontAnimate I2I · UFUC (unseen fonts, unseen contents) | SSIM: 0.7105 -> 0.7209 | Improved |
| add temporal part and frame-index (I2V) to I2I font model (ablation) | FontAnimate I2I · UFUC (unseen fonts, unseen contents) | PSNR: 12.08 -> 12.43 | Improved |
| add noise inversion condition alignment to I2V model (ablation) | FontAnimate I2V · UFUC (unseen fonts, unseen contents) | LPIPS: 0.1340 -> 0.1322 | Improved |
| add noise inversion condition alignment to I2V model (ablation) | FontAnimate I2V · UFUC (unseen fonts, unseen contents) | RMSE: 0.2483 -> 0.2470 | Improved |
| use full temporal self-attention instead of causal (ablation) | FontAnimate I2V · UFUC (unseen fonts, unseen contents) | RMSE: 0.2499 -> 0.2483 | Improved |
| use full temporal self-attention instead of causal (ablation) | FontAnimate I2V · UFUC (unseen fonts, unseen contents) | PSNR: 12.38 -> 12.43 | Improved |
| add zero-convolution module for source image latent (ablation) | FontAnimate I2V · UFUC (unseen fonts, unseen contents) | RMSE: 0.2483 -> 0.2504 | No clear effect |
| set noise inversion updating rate to 1e-3 instead of 1e-1 (ablation) | FontAnimate NI · UFUC (unseen fonts, unseen contents) | RMSE: 0.2920 -> 0.2470 | Improved |
| set noise inversion updating rate to 1e-3 instead of 1e-5 (ablation) | FontAnimate NI · UFUC (unseen fonts, unseen contents) | LPIPS: 0.1341 -> 0.1322 | Improved |
| FontAnimate vs second-best on seen contents | FontAnimate · UFSC (unseen fonts, seen contents) | LPIPS: 0.1348 -> 0.1211 | Improved |
Closest papers by meaning
- Generate Like Experts: Multi-Stage Font Generation by Incorporating Font Transfer Process into Diffusion Models (CVPR 2024 · similarity 0.92)
- Neural Transformation Fields for Arbitrary-Styled Font Generation (CVPR 2023 · similarity 0.91)
- FontDiffuser: One-Shot Font Generation via Denoising Diffusion with Multi-Scale Content Aggregation and Style Contrastive Learning (AAAI 2024 · similarity 0.90)
- IF-Font: Ideographic Description Sequence-Following Font Generation (NeurIPS 2024 · similarity 0.89)
- VQ-FONT: Few-Shot Font Generation with Structure-Aware Enhancement and Quantization (AAAI 2024 · similarity 0.88)
- DeepCalliFont: Few-Shot Chinese Calligraphy Font Synthesis by Integrating Dual-Modality Generative Models (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.
