Machine Learning and Type Design 03: “Typeface Quality Assurance Using Machine Learning”

*This article is created based on a presentation given at ATypI 2019 Tokyo.
(https://www.youtube.com/watch?v=va0NnxdgrqA)

In addition to human checking, it is believed that checking using machine learning can be done for quality assurance of typeface design. It is believed that the characters with consistent typeface have “typeface features” common to all characters. Constant quality can be maintained mechanically by finding these typeface features using machine learning and using this for design checking.

When this method is established, a check can be performed to maintain a consistent typeface even when the characters are designed by multiple designers. Also, when there are several design plans for one character, the better suited plan for the character can be selected mechanically.

There are other combinations of machine learning and typefaces, such as utilization in kerning and typesetting, apart from those introduced today. As this technology is expected to be utilized to a wide scope, attention must be paid to any relevant developments.

(HT)

Series archive Other / Machine Learning and Type Design