/ComputerVision

Implementations of algorithms and operations used in Computer Vision: solely for Neural style transfer.

Primary LanguageJupyter Notebook

ComputerVision

This repository only contains implementations for Neural styel transfer.

Neural Style Transfer

To use the Neural Style transfer implementation:

Run in Google Colab

from NeuralStyleTransfer import implementNTS as NST

set the image dimension and train and generate an image

NST.setImageDim(400,300)
NST.run(num_iteration=1000)

the setImageDim function in the module implementNTS takes 2 arguments:

width: the width of the images used
height: the height of the images

the run function in the module implementNTS takes 3 arguments:

num_iteration: number of iterations to train for
content_image: the path to the content image of size: (WIDTH=400, HEIGHT=300)
style_image: the path to the style image of size: (WIDTH=400, HEIGHT=300)

Requirement

  • Tensorflow
  • numpy
  • scipy