TextileNet: Towards Zero-shot Text-style Segmentation of Manuscripts

2026-07-10

AI Summary

This study addresses the challenge of automatic writer identification in historical manuscripts, where scarce annotations, open-set writer distributions, and image degradation hinder conventional approaches. To overcome these limitations, the authors propose the first zero-shot text style segmentation method, employing a fully convolutional multi-task network trained exclusively on synthetic data to generate pixel-level texture embeddings that enable zero-shot transfer to real historical manuscripts. The work also introduces a visual benchmark comprising 80 paleographic questions to establish human performance baselines in identifying late medieval scripts. Experimental results demonstrate that the model effectively supports zero-shot retrieval of both handwriting styles and author gender, validating its reliability in paleographic analysis while cautioning against overinterpretation of gender attribution.

Abstract

Automatic writer identification systems have progressed remarkably in recent years, yet their deployment in archival paleography remains limited by the scarcity of labeled training data, open scribe sets, and degraded image quality. We present TextileNet, a fully convolutional multi-task network trained exclusively on synthetic data to produce dense pixel-level texture embeddings, which we transfer zeroshot to historical manuscript analysis. As an original contribution to evaluation methodology, we designed a paleographic visual quiz of 80 pair and triplet questions and administered it to a range from lay participants to senior paleographers under strict anonymity, establishing to our knowledge for the first time a human baseline for script-style discrimination on late medieval text. We employ TextileNet embeddings to perform zero-shot retrieval on sub-word granularity for hand and gender identification. Our experimental results help in building the credibility of TextileNet in the paleographic domain, but more than that demonstrate in experimental terms that the question of gender in handwriting needs to be treated with caution.

Problem

• writer identification
• labeled data scarcity
• degraded manuscript images
• gender attribution in handwriting
• paleography

Innovation

• zero-shot segmentation
• synthetic data training
• texture embeddings
• paleographic analysis
• handwriting gender identification