Overview
The past has had a strong impact on letter drawing, as historical scripts, calligraphy styles, and typographic developments continue to shape modern design. Letterforms rooted in older practices reflect the culture, symbolism, artistic skills, and available tools of their times. From handwritten manuscripts to early printed texts, these traditions guide how modern lettering is structured and styled. Nowadays, designers often look to these older techniques for inspiration, blending classic elements with new, modern ideas.
The proposed project aims to help type designers better understand the influence of historical types on their current letterforms. By utilizing a set of digitized specimens (see an example here), the project seeks to train an algorithm that takes the designer’s letter drawing as input and provides a list of similar shapes from the specimens, along with explanations for why each was selected as similar.
This research project is part of a larger initiative led by the ECAL Master Type Design Department, under the principal investigation of Alice Savoie, a renowned French type designer and teacher, and Kai Bernau, a type designer, teacher, and founder of Atelier Carvalho Bernau. The participating student will be part of the EPFL+ECAL Lab team conducting the design research project, which is composed of engineers, designers, and psychologists. Ultimately, the created model will serve as a teaching tool for future type designers.
Key Machine Learning Domains
- Segment Anything Model (SAM)
Prerequisites
- Working knowledge of computer vision and machine learning models
- Proficiency in Python