Chinese font style migration method

Abstract

A Chinese font style migration method is characterized in that unsupervised style migration learning is carried out on handwritten Chinese characters, and two auxiliary networks are creatively added to a generator composed of cyclic generative adversarial networks on the basis of an original adversarial generative network. The method comprises the following steps: simultaneously extracting font structure features of an original image and a generator generation image through a Chinese character classification identification residual network, and taking matrix information of a specific layer as an auxiliary loss function to control the font structure consistency of the two images; and using the style encoder for extracting style characteristics of the graph generated by the generator and feeding back the style characteristics to the generator so as to control the style consistency of the two.

Description

technical field [0001] The invention relates to a Chinese font style transfer method. [0002] technical background [0003] Writing is a tool that people use to convey information and is an important part of human culture. The invention of Chinese characters not only enables the inheritance of the long history of our country, but also a wide variety of Chinese character fonts (such as the calligraphic style in the works of ancient literati and writers, and the artistic style designed by designers) play a role in visual communication and are an artistic symbol. . [0004] However, there have always been some pain points in the long-established field of Chinese character design, making the expansion of Chinese character fonts very difficult. First of all, when designing fonts, in order to ensure the uniformity of the glyphs, designers often spend a lot of time on the design of each character. At the same time, compared with English, which has only 26 letters, Chinese charac...

AI Technical Summary

Problems solved by technology

[0020] The first main shortcoming of CN111667008A technology is that it needs the Chinese characters of standard font and personalized Chinese character font to appear in pairs, namely Chinese character designer provides and selects the same Chinese character in the standard font data set, needs to design these Chinese characters word by word

Therefore, if the standard Chinese characters selected in the data set are changed, these designed Chinese characters also need to be replaced, which is still too expensive for the designer

[0021] The second disadvantage is that this technology can only generate one target font for the same source font. If you want to modify the generated personalized Chinese character style, you need a new round of design and training, so that the model obtained under this technology The scope of application is limited, the generation of personalized Chinese characters is not flexible enough, and there is no diversity, and the model cannot be used universally

Images