/DeepLearningandArt

neural style transfer

Primary LanguageJupyter Notebook

DeepLearningandArt

neural style transfer this model is to optimize algorithm updates the pixel values rather than the nueral entwork's parameters.

News

20180428 READ

Summary

  1. Neural Style Transfer is an algorithm that given a content image C and a style image S can generate an artistic image
  2. It uses representations (hidden layer activations) based on a pretrained ConvNet.
  3. The content cost function is computed using one hidden layer's activations.
  4. The style cost function for one layer is computed using the Gram matrix of that layer's activations. The overall style cost function is obtained using several hidden layers.
  5. Optimizing the total cost function results in synthesizing new images.