Skip to content

Neural Style Transfer to engineer images via Transfer Learning and Feature Extraction

Notifications You must be signed in to change notification settings

SimranAnand1/Neural-Style-Transfer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Neural-Style-Transfer

Style transfer consists in generating an image with the same "content" as a base image, but with the "style" of a different picture (typically artistic). This is achieved through the optimization of a loss function that has 3 components: "style loss", "content loss", and "total variation loss":

The total variation loss imposes local spatial continuity between the pixels of the combination image, giving it visual coherence. The style loss is where the deep learning keeps in --that one is defined using a deep convolutional neural network. Precisely, it consists in a sum of L2 distances between the Gram matrices of the representations of the base image and the style reference image, extracted from different layers of a convnet (trained on ImageNet). The general idea is to capture color/texture information at different spatial scales (fairly large scales --defined by the depth of the layer considered). The content loss is a L2 distance between the features of the base image (extracted from a deep layer) and the features of the combination image, keeping the generated image close enough to the original one..

About

Neural Style Transfer to engineer images via Transfer Learning and Feature Extraction

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published