Source

This notebook is base on a blogpost by Aman Kumar Mallik that can be found here:

link text

#Neural Style Transfer Using PyTorch

###Introduction

Neural style transfer is an artificial system based on the Deep Neural Network to generate artistic images. This approach uses two random images, the content and the style image. It extracts the structural features from the content image, whereas the style features from the style image.

#Content and Style Representation

###Content Representation

The convolutional neural network develops representations of the image along the processing hierarchy. As we move deeper into the network, the representations will care more about the structural features or the actual content than the detailed pixel data. To obtain these representations, we can reconstruct the images using the feature maps of that layer. Reconstruction from the lower layer will reproduce the exact image. In contrast, the higher layer's reconstruction will capture the high-level content and hence we refer to the feature responses from the higher layer as the content representation.