Research map / Image generation / Text-to-image diffusion
Handwriting Transformers
Ankan Kumar Bhunia, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Fahad Shahbaz Khan, Mubarak Shah
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add transformer encoder to the baseline (ablation) | HWT baseline · IAM OOV-U | FID: 134.45 -> 128.80 | Improved |
| add transformer decoder to the baseline (ablation) | HWT baseline · IAM OOV-U | FID: 134.45 -> 124.81 | Improved |
| add both transformer encoder and decoder to the baseline (ablation) | HWT baseline · IAM OOV-U | FID: 134.45 -> 116.50 | Improved |
| add cycle loss to encoder-decoder transformer (ablation) | HWT · IAM OOV-U | FID: 116.50 -> 114.10 | Improved |
| use character-level instead of word-level content conditioning (ablation) | HWT · IAM OOV-U | FID: 126.87 -> 114.10 | Improved |
| use HWT instead of GANwriting for unseen words and unseen styles | styled generation · IAM OOV-U | FID: 130.68 -> 114.10 | Improved |
| use HWT instead of ScrabbleGAN and Davis et al. for unconditioned generation quality | handwritten text generation · IAM handwritten text generation | FID and GS: 20.65 -> 19.40 | Improved |
| use HWT instead of GANwriting and Davis et al. for human style mimicry | styled generation · User preference study | preference responses: 810 out of 1000 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GANwriting — GANwriting: Content-Conditioned Generation of Styled Handwritten Word Images (ECCV 2020) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning to Generate Stylized Handwritten Text via a Unified Representation of Style, Content, and Noise (ICLR 2026 · similarity 0.89)
- Zero-Shot Styled Text Image Generation, but Make It Autoregressive (CVPR 2025 · similarity 0.88)
- Handwritten Text Generation From Visual Archetypes (CVPR 2023 · similarity 0.88)
- Disentangling Writer and Character Styles for Handwriting Generation (CVPR 2023 · similarity 0.88)
- DreamText: High Fidelity Scene Text Synthesis (CVPR 2025 · similarity 0.84)
- Decoupling Layout from Glyph in Online Chinese Handwriting Generation (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
