Research map / 3D vision / Face and shape
Joint Implicit Neural Representation for High-fidelity and Compact Vector Fonts
Chia-Hao Chen, Ying-Tian Liu, Zhifei Zhang, Yuan-Chen Guo, Song-Hai Zhang
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 probabilistic corner field to the SDF (ablation) | joint implicit font representation · Font reconstruction | L1: .0160 -> .0140 (128^2) | Improved |
| add peak constraint loss on corner field (ablation) | corner field · Corner count error | L1 error of corner number: .6832 -> .5806 (level 0.5) | Improved |
| add signed distance flow loss to align CF and SDF (ablation) | corner field interpolation · Font interpolation | corner divergence: 971.7 -> 988.6 | Improved |
| add HyperNetworks to SDF network (ablation) | SDF network · Font reconstruction 512^2 | L1: .0455 -> .0154 | Improved |
| increase HyperNetwork parameters (3 to 4 layers) (ablation) | SDF network · Font reconstruction 512^2 | L1: .0151 vs .0154 | No clear effect |
| use method vs DeepVecFont on interpolation L1 | Ours · Font interpolation | L1: .0485 vs .0454 (DeepVecFont) | Mixed |
| use implicit-field vectorization instead of sequence generation | vector font output · Compactness | commands and control points: CMDs 32.25 vs 22.75 (DeepVecFont) | Mixed |
Closest papers by meaning
- VecFontSDF: Learning To Reconstruct and Synthesize High-Quality Vector Fonts via Signed Distance Functions (CVPR 2023 · similarity 0.91)
- A Multi-Implicit Neural Representation for Fonts (NeurIPS 2021 · similarity 0.89)
- VecFusion: Vector Font Generation with Diffusion (CVPR 2024 · similarity 0.87)
- VecGlypher: Unified Vector Glyph Generation with Language Models (CVPR 2026 · similarity 0.85)
- Vector Grimoire: Codebook-based Shape Generation under Raster Image Supervision (ICML 2025 · similarity 0.85)
- Towards High-fidelity Artistic Image Vectorization via Texture-Encapsulated Shape Parameterization (CVPR 2024 · similarity 0.83)
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.
