Embedding Font Impression Word Tags Based on Co-occurrence
This work addresses font generation and retrieval for designers or users seeking impression-based typography, but it is incremental as it builds on existing embedding techniques.
The paper tackled the problem of embedding font impression tags by leveraging shape-impression relationships, proposing a novel method based on co-occurrence graphs and spectral embedding, which outperformed BERT and CLIP in impression-guided font generation.
Different font styles (i.e., font shapes) convey distinct impressions, indicating a close relationship between font shapes and word tags describing those impressions. This paper proposes a novel embedding method for impression tags that leverages these shape-impression relationships. For instance, our method assigns similar vectors to impression tags that frequently co-occur in order to represent impressions of fonts, whereas standard word embedding methods (e.g., BERT and CLIP) yield very different vectors. This property is particularly useful for impression-based font generation and font retrieval. Technically, we construct a graph whose nodes represent impression tags and whose edges encode co-occurrence relationships. Then, we apply spectral embedding to obtain the impression vectors for each tag. We compare our method with BERT and CLIP in qualitative and quantitative evaluations, demonstrating that our approach performs better in impression-guided font generation.