Unlock instant, AI-driven research and patent intelligence for your innovation.

Personalized font generation method based on CycleGAN

A font and personalized technology, applied in neural learning methods, electrical digital data processing, biological neural network models, etc., can solve problems such as narrow scope of application, difficulty in obtaining paired images, etc., achieve high clarity and consistency, easy acquisition, Generating image-realistic effects

Pending Publication Date: 2022-03-01
WUHAN UNIV
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This model borrows the idea of ​​style transfer from ordinary images, realizes the style transfer from Chinese character images to Chinese character images, and completes the task of Chinese character image style transfer. Ordinary deep learning networks need a large number of paired images for training, but some paired images are more difficult to obtain. Difficult, narrow scope of application

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Personalized font generation method based on CycleGAN
  • Personalized font generation method based on CycleGAN
  • Personalized font generation method based on CycleGAN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] The present invention provides a kind of personalized font generation method based on CycleGAN, and research content includes the following several parts:

[0046] (1) Use the Song-style Chinese character dataset and the self-built handwritten Chinese character dataset as the training set, and adopt a data enhancement method to invert the Chinese characters to double the number of datasets, and ensure that the Song-style and handwritten content are consistent.

[0047] (2) Construct a Cycle Generative Adversarial Network (CycleGAN) to realize the style transfer of Chinese character images. Input the pictures in the preprocessed training (train) data set for model training, and use the data in the test set (test) for testing to observe the effect of font style transfer. On this basis, the parameters, loss function and generator structure of the model are improved to output better Chinese character image style transfer pictures.

[0048] The construction steps of the cyc...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a personalized font generation method based on a CycleGAN (Graphics Area Network). According to the method, starting from the idea of image style migration, a model for font style migration is established based on a cyclic generative adversarial network (CycleGAN), and the main innovation point is that a picture style migration method is applied to font style design, a generator of the cyclic generative adversarial network is optimized, and the font style migration model is applied to font style design. An original Resnet structure in a generator is replaced by a U-net structure, proper model parameters are selected, a data set belonging to the generator is constructed, data enhancement operation is carried out on the data set, and finally a generative style font picture with a good effect is obtained.

Description

technical field [0001] The invention belongs to the field of style image generation, and in particular relates to a font image generation model based on a cyclic generation confrontation network for handwritten personalized fonts. Background technique [0002] Chinese font style design refers to directly converting the font of the target content into a style font based on a small number of existing style fonts. As a very complex language for expressing meaning, the design of Chinese fonts is extremely complicated. The number and types of Chinese characters are far more than the English letters with phonetic representations. Therefore, it takes a huge amount of work to complete the font design of Chinese characters with a relatively complete coverage. In addition, in the era of pursuing personalization, many people require the establishment of personal font libraries. However, the establishment of a font library requires individuals to write and upload a large number of Chi...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F40/109G06N3/04G06N3/08
CPCG06F40/109G06N3/08G06N3/045
Inventor 李治江徐慧婷黄煜城曹丽琴
Owner WUHAN UNIV