/Deep-Learning-Final-Project

Primary LanguagePython

ECBM E4040 Fall 2017 Final Project

Project Title: A Neural Algorithm of Artistic Style

  • Term: Fall 2017

  • Team ID: AAAA

Team Members

  • Qingyuan Zhang qz2283
  • Yike Peng yp2425
  • Renjie Li rl2928

Link to original paper: https://arxiv.org/pdf/1508.06576v2.pdf

File Description:

  • neural_style.py : main function to achieve content and style inputting, image synthesis, optimization and synthetic image generation

  • synthesis.py : generate images which mix content with style

  • style_layer_cal.py : calculate style layer weights

  • feature_cal.py : extract features of content and style in certain layers of CNN

  • total_loss.py : compute total loss function which consists of content loss and style loss, and implement noise-reduce optimization

  • vgg.py : provide VGG19-Networks

Outlook

Output Ratio

Note

To reproduce our work, clone the whole repository onto your own local machine. In your terminal, run this code:

python neural_style.py --content Pics/Content/boston.jpg --styles Pics/Styles/composition_vii.jpg --iterations 500 --output [your output path and output name]

Please make sure you download the pre-trained VGG19 mat model into the pretrained_VGG19 folder, and then run above code.

Contribution statement:

All team members contributed equally in all stages of this project. All team members approve our work presented in this repository including this contributions statement.