NIV: Neural Axis Variations for Variable Font Generation
Researchers have developed NIV (Neural Axis Variations), a novel method to automatically convert static fonts into variable fonts. This technique predicts per-point displacements on vector glyph geometry, capturing complex interactions between design axes like weight and width. Trained on a large dataset derived from variable Google Fonts, NIV demonstrates generalization across various glyph types, including complex CJK characters and handwriting, producing standard variable font files. AI
IMPACT This method could streamline font design and enable more dynamic typographic applications.