Emotion-based Hangul font recommendation system using crowdsourcing

Emotion analysis of text documents is an emerging area of interest, and it is closely related to the visualization of emotions and sentiments. There is a need to consider a different setting for emotion visualization of text through fonts since typography forms a huge part of visualization. Several studies were conducted on techniques that use content expressions that match the context and the font emotion, including textual sentiment in combination. However, studies on emotion with regard to the Korean Hangul fonts have thus far not yielded a consensus. Therefore, we propose a Hangul font recommendation system that uses emotion classification properties extended from a World Wide Web Consortium (W3C) standard and is based on a Maximum likelihood estimation approach with crowdsourced attribute values. We verified the reliability and validity of the emotion attributes as applied to Hangul fonts and compared the efficiency of shape-, purpose-, and emotion-based recommendations. We show that users can find the appropriate font based on emotion recognition in a commercial font set using this system.