Handwritten calligraphy character generation method based on GAN network

Abstract

The invention provides a handwritten calligraphy character generation method based on a GAN network, belongs to the technical field of handwritten generation, deep learning and neural networks, and aims to collect calligraphy characters into images, extract features of the images and complete generation of handwritten calligraphy character images based on styles and Chinese character contents through a generative adversarial network. The whole GAN network model is composed of a single word generation network and a page word generation network; a generator and a discriminator are alternately trained through a GAN network, learning of a single-character network is completed, the relation between single characters is learned through an LSTM network, and a final page handwritten calligraphy character network model is formed and used for generation of designated character calligraphy.

Description

technical field [0001] The invention relates to handwriting generation, deep learning and neural network technology, and in particular to a GAN network-based handwriting calligraphy character generation method. Background technique [0002] Generative Adversarial Networks (GAN, Generative Adversarial Networks) is a deep learning model, originally proposed by Ian Goodfellow, and is one of the most important methods for unsupervised learning on complex distributions in recent years. The GAN model produces high-quality output through the mutual game learning of the two modules in the framework, the Generator (Generator) and the Discriminator (Discriminator). The goal is to train a generative model to perfectly fit the real data distribution so that the discriminative model cannot be distinguished. The role of the generative model is to simulate the distribution of real data. The role of the discriminant model is to judge whether a sample is a real sample or a generated sample. ...

AI Technical Summary

Problems solved by technology

This kind of handwritten characters are of different sizes and shapes. It is difficult to achieve a well-proportioned effect in the computer font library, so it can only be passed down by handwriting.

Images