Character level attention siamese networks for Chinese handwriting comparison
Li Zhu, Long Tang, Huaqing Mao · 2026
Read Character level attention siamese networks for Chinese handwriting comparison by Li Zhu, Long Tang, Huaqing Mao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353503.txt.
Description
Li Zhu, Long Tang, and Huaqing Mao published a study on machine reading of Chinese handwriting, in PLOS ONE in 2026. Judging whether two samples were written by the same person is hard in Chinese, with thousands of characters and wide personal style. Their system compares samples at the level of single characters.
Zhu's team built paired networks with an attention step, so the model focuses on the strokes that matter most. Trained this way, it learns what makes writing samples match or differ. Forensic examiners and archivists could use such a tool as a tireless second opinion on disputed documents. It would not replace human judgment, but it can sort large sets of samples quickly.
Catalog record
| CREATOR | Li Zhu, Long Tang, Huaqing Mao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50497 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |