“The universe of fonts, charted by machine” by Loviscach

  • ©Jörn Loviscach




    The universe of fonts, charted by machine



    Today, computer fonts come in hundreds and thousands. How do you find a font that fits for the task at hand? Which general types are available? Which ones are related—or not related at all—to the ones you know well? To address such issues, this work applies strategies from Music Information Retrieval (MIR) to organizing fonts. Finding similar music and graphically laying out a music collection according to the similarity of the tracks are standard tasks in MIR, tackled by automatically extracting meaningful but low-level descriptors from the bare content data—that is: waveforms—and discovering high-level meaning through machine learning.

ACM Digital Library Publication:

Overview Page: