Skip to content

Repository files navigation

Style Transfer with improved Color Consistency Loss.

Color Consistency Loss

Color Consistency Loss essentially tries to minize the Eulidean Distance between the IQ dimension in the YIQ space between the style image and the generated image, therefore allowing the network to selective choose which areas to apply styles and which areas to avoid applying styles. As a result, network with color consistency loss tends to produce less noisy and more visually appealing results.

Starry Night

StarryNight

Without Color Consistency Loss: UR_StarryNight_Original

With Color Consistency Loss: UR_StarryNight_Improved

Without Color Consistency Loss: Zheng_StarryNight_Original

With Color Consistency Loss: Zheng_StarryNight_Improved

Van Gogh

Vangogh

Without Color Consistency Loss: UR_Vangogh_Original

With Color Consistency Loss: UR_Vangogh_Improved

Training Instruction to be updated...

About

Neural Style Transfer

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages