Method for guiding Chinese character font generation based on structure

Abstract

The invention discloses a method for guiding Chinese character font generation based on a structure. The method comprises the following steps of 1, preparing font image data; 2, extracting the outlineand the skeleton of the font image; 3, constructing a generative adversarial network by adopting an encoder and decoder structure as a generator and a convolutional neural network as a discriminator;4, training the constructed neural network by utilizing the standard font picture, the outline, the skeleton and the target font picture data set; 4, inputting the complete standard font into a generator of the trained adversarial generative network. The generative adversarial network is trained through a small number of target fonts, and the complete target font library file can be obtained. Inthis way, the neural network training problem is easier to handle, and the outline and the structure of the font are more natural.

Description

technical field [0001] The invention relates to a method for generating Chinese character fonts, specifically a method for generating Chinese character fonts based on structure guidance. Background technique [0002] With the rise of the Internet, more and more people now like to use personalized fonts in different situations. On the one hand, different font styles can flexibly express individuality, endowing the text with the unique characteristics and signs of the writer. On the other hand, the glyphs in the personalized font style bring dynamic visual perception and can attract more attention in various social network media. However, creating a handwritten Chinese font library is still a time-consuming and laborious work with a long production cycle. The handwritten fonts available on electronic devices are mainly produced by professional font designers in commercial companies. The design of the font library is a very time-consuming job. Unlike the English font library ...

AI Technical Summary

Problems solved by technology

Furthermore, for characters with complex structures and or cursive handwriting styles, these end-to-end methods often produce results with unreasonable strokes or incorrect structures

Images