Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.21098 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915468610109440 |
|---|---|
| author | Jin, Zezhong Desai, Shubhang Chen, Xu Fang, Biyi Huang, Zhuoyi Li, Zhe Gan, Chong-Xin Tu, Xiao Mak, Man-Wai Lu, Yan Liu, Shujie |
| author_facet | Jin, Zezhong Desai, Shubhang Chen, Xu Fang, Biyi Huang, Zhuoyi Li, Zhe Gan, Chong-Xin Tu, Xiao Mak, Man-Wai Lu, Yan Liu, Shujie |
| contents | In this paper, we propose TrInk, a Transformer-based model for ink generation, which effectively captures global dependencies. To better facilitate the alignment between the input text and generated stroke points, we introduce scaled positional embeddings and a Gaussian memory mask in the cross-attention module. Additionally, we design both subjective and objective evaluation pipelines to comprehensively assess the legibility and style consistency of the generated handwriting. Experiments demonstrate that our Transformer-based model achieves a 35.56\% reduction in character error rate (CER) and an 29.66% reduction in word error rate (WER) on the IAM-OnDB dataset compared to previous methods. We provide an demo page with handwriting samples from TrInk and baseline models at: https://akahello-a11y.github.io/trink-demo/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_21098 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TrInk: Ink Generation with Transformer Network Jin, Zezhong Desai, Shubhang Chen, Xu Fang, Biyi Huang, Zhuoyi Li, Zhe Gan, Chong-Xin Tu, Xiao Mak, Man-Wai Lu, Yan Liu, Shujie Computation and Language Artificial Intelligence In this paper, we propose TrInk, a Transformer-based model for ink generation, which effectively captures global dependencies. To better facilitate the alignment between the input text and generated stroke points, we introduce scaled positional embeddings and a Gaussian memory mask in the cross-attention module. Additionally, we design both subjective and objective evaluation pipelines to comprehensively assess the legibility and style consistency of the generated handwriting. Experiments demonstrate that our Transformer-based model achieves a 35.56\% reduction in character error rate (CER) and an 29.66% reduction in word error rate (WER) on the IAM-OnDB dataset compared to previous methods. We provide an demo page with handwriting samples from TrInk and baseline models at: https://akahello-a11y.github.io/trink-demo/ |
| title | TrInk: Ink Generation with Transformer Network |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2508.21098 |