DNA: Dual-branch Network with Adaptation for Open-Set Online Handwriting Generation

Online handwriting generation (OHG) enhances hand- writing recognition models by synthesizing diverse, human- like samples. However, existing OHG methods strug- gle to generate unseen characters, particularly in glyph- based languages like Chinese, limiting their real-world ap- plicability. In this paper, we introduce Open-set Online Handwriting Generation (OOHG). In this new task, the writer’s style and the characters generated during testing are unseen during training. To tackle this challenge, we pro- pose a Dual-branch Network with Adaptation (DNA), which comprises an adaptive style branch and an adaptive con- tent branch. The style branch learns stroke attributes such as writing direction, spacing, placement, and flow to gen- erate realistic handwriting. Meanwhile, the content branch is designed to generalize effectively to unseen characters by decomposing character content into structural information and texture details, extracted via local and global encoders, respectively. Extensive experiments demonstrate that our DNA model is well-suited for the OOHG setting, achieving state-of-the-art performance.