Jun 01, 2021 Article blog
This article was reproduced to Know ID: Charles (Bai Lu) knows his personal column
Download the W3Cschool Mobile App, 0 Foundation Anytime, Anywhere Learning Programming >> Poke this to learn
This is probably one of the hottest applications of deep learning last year, using neural networks to migrate the world's most famous paintings to any ordinary picture.
Let's put the picture first!
On the left is the picture to be converted, and on the right is the picture for learning the style of painting:
outcome:
One more set:
Baidu web download link: https://pan.baidu.com/s/1yJTxYBmV2rfLdNt2s53N_A
Password: bqjs
The documents given include:
The principle paper on the implementation of the project: "A Neural Of Of The Style";
Source file for related projects on GitHub (TensorFlow implementation);
The Keras version of the project implementation (which is also the source code for the first two examples, I made some comments).
Instructions for use
STEP1:
Place the provided models folder under a path similar to the one shown in the following image:
STEP2:
Run the ImitationPaintings.py files provided, in the form of python ImitationPaintings.py, the path to be converted, the template picture path, the suffoating path plus the name of the saved path for the generated picture (note that there is no need to have a suffix such as .jpg), for example:
Implementation is so simple, the default number of iterations is 10, running a little longer, as for the specific principles, you can see the paper and related GitHub projects (because the theoretical basis required is too much, I do not know how to introduce its principles, here recommend a blog post: http://www.cnblogs.com/mangoyuan/p/6329410.html personally feel written in general, want to understand the specific principles or need to study the paper and source code line).
1. Based on the implementation of python deep learning library DeepPy:
https://github.com/andersbll/neural_artistic_style
2. Based on the implementation of Python Deep Learning Library Caffe:
https://github.com/fzliu/style-transfer
3. Based on the implementation of python Deep Learning Library TensorFlow:
https://github.com/log0/neural-style-painting
Finally, put some more pictures to attract everyone
Here are some other people running out of the results, feel very nice.