Research map / Image generation / Text-to-image diffusion
Content and Style Aware Generation of Text-Line Images for Handwriting Recognition
Lei Kang, Pau Riba, Marçal Rusiñol, Alícia Fornés, Mauricio Villegas
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 generated samples (WikiText lexicon) to IAM training set | HTR recognizer · IAM test | CER (%): 10.46 -> 9.66 | Improved |
| use IAM lexicon instead of WikiText for generated text | HTR recognizer · IAM test | CER (%): 9.66 -> 9.37 | Improved |
| add data augmentation on real and generated samples | HTR recognizer · IAM test | CER (%): 9.37 -> 8.62 | Improved |
| add generated samples and augmentation (proposed vs ScrabbleGAN) | HTR recognizer · IAM test | CER (%): 13.57 -> 8.62 | Improved |
| adapt recognizer with synthetic samples mimicking unlabeled Rimes style | HTR recognizer · Rimes test | CER (%): 27.30 -> 18.19 | Improved |
| use Rimes-style instead of IAM-style synthetic adaptation data | HTR recognizer · Rimes test | CER (%): 20.55 -> 18.19 | Improved |
| use recognizer trained independently instead of jointly with generator | HTR recognizer · IAM test | CER (%): 16.73 -> 10.46 | Improved |
| replace VGG19 with ResNet34 convolutional layers (ablation) | generator · IAM test | vFID: 115.46 -> 114.39 | Improved |
| replace VGG19 with ResNet34 convolutional layers (ablation) | generator · IAM test | speed (ms): 144.13 -> 136.80 | Improved |
| replace Seq2Seq recognizer with Transformer recognizer (ablation) | generator · IAM text-lines (1-90 chars) | vFID: 249.94 -> 114.39 | Improved |
| replace Seq2Seq recognizer with Transformer recognizer (ablation) | generator · IAM words (1-10 chars) | vFID: 136.51 -> 146.74 | Hurt |
| add global string encoding to local character-wise encoding (ablation) | generator · IAM words (1-10 chars) | vFID: 162.38 -> 146.74 | Improved |
Closest papers by meaning
- Handwritten Text Generation From Visual Archetypes (CVPR 2023 · similarity 0.88)
- GANwriting: Content-Conditioned Generation of Styled Handwritten Word Images (ECCV 2020 · similarity 0.88)
- Zero-Shot Styled Text Image Generation, but Make It Autoregressive (CVPR 2025 · similarity 0.87)
- General Detection-based Text Line Recognition (NeurIPS 2024 · similarity 0.85)
- MetaWriter: Personalized Handwritten Text Recognition Using Meta-Learned Prompt Tuning (CVPR 2025 · similarity 0.85)
- 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.
