Compressed ideograph -visualized-
This work is a project to create Kanji (Chinese characters) by deep learning, which is different from the conventional method of creating Kanji. It uses a text2image based model to infer and visualize a compressed Kanji for an arbitrary character string or text input. This work attempts to give Kanji characters to sentences and character strings that have not been given specific Kanji characters. It also brings new interpretations to concepts that already have Kanji characters.