Description

Despite its potential for researchers and consumer platforms, end-to-end Handwritten Text Recognition (HDR) remains an underexplored field. Unlike traditional OCR, HDR recognizes entire documents, capturing both text and structural elements such as layout and reading order. Recent models, such as DAN, have demonstrated advancements in whole-document recognition. This project will apply HDR models to early modern Spanish manuscripts, leveraging transformer-based architectures to improve recognition accuracy. By integrating state-of-the-art machine learning with historical document analysis, we aim to enhance the accessibility of handwritten and printed texts from this period while contributing to the broader field of AI-driven humanities research.

Duration

Total project length: 175 hours

Task ideas

• Creation of an self-supervised AI model capable of performing text recognition

Expected results

• Machine learning models will be trained to perform text recognition of non-standard printed text
• AI should be able to extract text with at least 80% accuracy

Requirements

Python and some previous experience in Machine Learning.

Difficulty level

Advanced

Mentors

• Sergei Gleyzer (University of Alabama)
• Xabier Granja (University of Alabama)
• Emanuele Usai (University of Alabama)

Please DO NOT contact mentors directly by email. Instead, please email human-ai@cern.ch with Project Title and include your CV and test results. The mentors will then get in touch with you.

Corresponding Project

• RenAIssance

Participating Organizations

• Alabama
• Delhi

---