Skip to content

sfarhat/neural-style-transfer

Repository files navigation

Neural Style Transfer (+ Seam Carving)

This is an implementation of Gatys et al's seminal paper on Neurl Style Transfer, a method of extracing content and stylistic information from a Convolutional Neural Network to convert one image into the style of another. My implementation differs from the paper's in my choice of layers used, loss functions, and specifics of the descent step.

You can find the report detailing my choices and the paper itself, in addition to my results from implementing the Seam Carving algorithm as well, here