# Master-art-punk **Repository Path**: yangemail8/Master-art-punk ## Basic Information - **Project Name**: Master-art-punk - **Description**: 通过机器学习世界名画制作属于你的大师级艺术朋克头像 - **Primary Language**: Python - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 4 - **Created**: 2022-08-23 - **Last Updated**: 2022-08-23 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Master-art-punk ##### Make your master artistic punk avatar through machine learning world famous paintings. ##### 通过机器学习世界名画制作属于你的大师级艺术朋克头像 ##### Nowadays, NFT is becoming popular in the world, and various trading platforms like opensea have developed rapidly, and cryptopunks have even sold tens of millions of dollars. ##### 如今,NFT在世界上正流行起来,各种交易平台像opensea等,快速发展了起来,像cryptopunks甚至卖出了上千万美金。 ##### This project will help you to know in general how these NFT pictures are made, and try to make your own NFT. ![punk1](https://gitee.com/philipjhc/Master-art-punk/blob/main/Master-art-punk/cover_photo/example.png) ![punk2](https://gitee.com/philipjhc/Master-art-punk/blob/main/Master-art-punk/cover_photo/example2.png) ##### 这个项目将有助于你大体知道这些NFT的图片是如何制作的,并尝试制作自己的NFT。 ##### This project uses the K-Means algorithm to acquire and learn the color matching of world famous paintings in the data set, and imitate its color style to color the model. ##### 本项目运用了K-Means算法对数据集中世界名画的色彩搭配进行获取和学习,并模仿其色彩风格给模型上色,展示如下: ![NFT1](https://gitee.com/philipjhc/Master-art-punk/blob/main/Master-art-punk/cover_photo/NFT1.png) ![NFT1](https://gitee.com/philipjhc/Master-art-punk/blob/main/Master-art-punk/cover_photo/NFT4.png) ## 使用说明: ### 1.安装要求: #### 1.更改settings.py的第8行,填入你自己的图片训练集路径,例如,打开./data/monet看到图片类似。(类似方法制作数据集,默认png格式) #### 2.安装numpy和pypng:pip install numpy / pip install pypng ### 2.训练: #### 1.python model.py / 运行 model.py (等待时间跟图片数目有关) #### 2.运行后,可发现colordistance.csv更新(需修改名字在setting.py的color_distance_filepath修改) #### 3.接着打开app.py 第12行选择模组类型(内置模型可在imageData下的stickers.py,subject.py中查看),第10行设置生成数量,第68行设置输出文件路径。 #### 4.app.py #### 5.output对应文件夹下中有已生成的朋克头像。(最好选择一个作家的风格,例如,我选择了莫奈)