BoundInk: Boundary-Aware Online Handwriting Generation
BoundInk generates writer-conditioned online handwriting by modeling inter-character boundaries, improving joins, spacing, and human preference.
BoundInk is a writer-conditioned online handwriting generator that models boundaries between characters explicitly instead of relying only on long-range sequence modeling. It jointly models local transitions and surrounding text to preserve writer-specific glyphs while improving cursive joins and spacing. A boundary-aware evaluation measures continuity and spatial relationships beyond conventional trajectory similarity. Across three benchmark-matched settings it improves boundary-quality measures, reduces normalized dynamic time warping by 17.6-47.8%, and is preferred in 78.0-82.6% of valid blind human judgments.
- Treats inter-character boundaries as explicit handwriting generation units
- Reduces normalized dynamic time warping by 17.6-47.8%
- Preferred in 78.0-82.6% of blind human judgments
- Adds evaluation of cursive continuity and spacing beyond trajectory similarity
From the original article
1 min read at huggingface.co
Realistic online handwriting depends not only on individual character shapes, but also on how a writer connects, spaces, and aligns adjacent characters. Existing methods rely primarily on long-range sequence modeling and capture these inter-character behaviors only implicitly. This often produces plausible glyphs accompanied by broken cursive joins, inconsistent spacing, or writer-inconsistent transitions.
By Jinsu Shin, Sungeun Hong, JinYeong Bak · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.