Quo Vadis Handwritten Text Generation for Handwritten Text Recognition?

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Pippi, Vittorio, Nikolaidou, Konstantina, Cascianelli, Silvia, Retsinas, George, Sfikas, Giorgos, Cucchiara, Rita, Liwicki, Marcus
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915444698382336
author Pippi, Vittorio
Nikolaidou, Konstantina
Cascianelli, Silvia
Retsinas, George
Sfikas, Giorgos
Cucchiara, Rita
Liwicki, Marcus
author_facet Pippi, Vittorio
Nikolaidou, Konstantina
Cascianelli, Silvia
Retsinas, George
Sfikas, Giorgos
Cucchiara, Rita
Liwicki, Marcus
contents The digitization of historical manuscripts presents significant challenges for Handwritten Text Recognition (HTR) systems, particularly when dealing with small, author-specific collections that diverge from the training data distributions. Handwritten Text Generation (HTG) techniques, which generate synthetic data tailored to specific handwriting styles, offer a promising solution to address these challenges. However, the effectiveness of various HTG models in enhancing HTR performance, especially in low-resource transcription settings, has not been thoroughly evaluated. In this work, we systematically compare three state-of-the-art styled HTG models (representing the generative adversarial, diffusion, and autoregressive paradigms for HTG) to assess their impact on HTR fine-tuning. We analyze how visual and linguistic characteristics of synthetic data influence fine-tuning outcomes and provide quantitative guidelines for selecting the most effective HTG model. The results of our analysis provide insights into the current capabilities of HTG methods and highlight key areas for further improvement in their application to low-resource HTR.
format Preprint
id arxiv_https___arxiv_org_abs_2508_09936
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Quo Vadis Handwritten Text Generation for Handwritten Text Recognition?
Pippi, Vittorio
Nikolaidou, Konstantina
Cascianelli, Silvia
Retsinas, George
Sfikas, Giorgos
Cucchiara, Rita
Liwicki, Marcus
Computer Vision and Pattern Recognition
Digital Libraries
The digitization of historical manuscripts presents significant challenges for Handwritten Text Recognition (HTR) systems, particularly when dealing with small, author-specific collections that diverge from the training data distributions. Handwritten Text Generation (HTG) techniques, which generate synthetic data tailored to specific handwriting styles, offer a promising solution to address these challenges. However, the effectiveness of various HTG models in enhancing HTR performance, especially in low-resource transcription settings, has not been thoroughly evaluated. In this work, we systematically compare three state-of-the-art styled HTG models (representing the generative adversarial, diffusion, and autoregressive paradigms for HTG) to assess their impact on HTR fine-tuning. We analyze how visual and linguistic characteristics of synthetic data influence fine-tuning outcomes and provide quantitative guidelines for selecting the most effective HTG model. The results of our analysis provide insights into the current capabilities of HTG methods and highlight key areas for further improvement in their application to low-resource HTR.
title Quo Vadis Handwritten Text Generation for Handwritten Text Recognition?
topic Computer Vision and Pattern Recognition
Digital Libraries
url https://arxiv.org/abs/2508.09936