Skip to content

chenmingTHU/image_style_transfer

Repository files navigation

Image Style Transfer

一个图像风格转换软件

An image style transfer software

小组成员:刘昊伟(组长),陈铭,刘晓明,许立,张舒楠

Members: Howie Lau, Ming Chen, Xiaoming Liu, Li Xu, Shunan Zhang

简介

这是一个转换图像风格的软件,主要提供以下两种转换风格:

  • 预设风格转换
  • 任意风格转换

预设风格转换类似滤镜,可以将图片转换为预先设定好的某种风格,时间较快。

任意风格转换则需要用户自定义输入风格图片,然后将内容图片转换为风格图片上的风格,时间较慢。

Introduction

This software can change the style of an image with 2 main functions:

  • Predefined style transfer
  • Arbitary style transfer

Predefined style transfer is similar to image filter, which is changing the image style through a pre-trained model with fast speed.

Arbitary style transfer need the user to specify the style image, and transfer the content image with lower speed.

使用

本程序暂只支持Python3。

受GitHub文件上传大小限制,无法将预训练好的模型上传到这里。

完整软件大小约2GB,同时我们也提供了只有任意风格转换的程序,大小约1GB。

百度网盘链接: https://pan.baidu.com/s/1kVSqt3x 密码:snpq

直接下载后运行。

cd /path/to/image_style_transfer

python3 MainWindow.py

注意:使用本软件需要安装PyQt5TensorflowPillow

暂未测试Linux的使用

在Windows下使用Python编译器(如PyCharm)直接运行MainWindow.py即可。

若要在Windows上使用,需要将PreDefined.py中179行和UserDefined.py中118行命令的python3改为python,同时将MainWindow.py第40行改为:

with open("template.qss", encoding="UTF-8") as file:

Usage

This software only supports Python3 for now.

Limited by the file storage of GitHub, we can't upload the pre-trained model here.

The whole package is about 2GB, and we also provide arbitary-only package, which is about 1GB.

Baidu Netdisk Link: https://pan.baidu.com/s/1kVSqt3x Password: snpq

cd /path/to/image_style_transfer

python3 MainWindow.py

Dependencies: PyQt5, Tensorflow, Pillow

Usage on Linux is not tested yet

If used on Windows, execute MainWindow.py with Python complier (PyCharm, etc.).

If you are using it on Windows, you need to change the command python3 to python on line 179 of PreDefined.py and line 118 of UserDefined.py, and change line 40 of MainWindow.py to:

with open("template.qss", encoding="UTF-8") as file:

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages