A Sliced Wasserstein Loss for Neural Texture Synthesis

Eric Heitz; Kenneth Vanhoey; Thomas Chambon; Laurent Belcour

CVPR 2021poster77 citations

A Sliced Wasserstein Loss for Neural Texture Synthesis

Eric Heitz, Kenneth Vanhoey, Thomas Chambon, Laurent Belcour

Abstract

We address the problem of computing a textural loss based on the statistics extracted from the feature activations of a convolutional neural network optimized for object recognition (e.g. VGG-19). The underlying mathematical problem is the measure of the distance between two distributions in feature space. The Gram-matrix loss is the ubiquitous approximation for this problem but it is subject to several shortcomings. Our goal is to promote the Sliced Wasserstein Distance as a replacement for it. It is theoretically proven, practical, simple to implement, and achieves results that are visually superior for texture synthesis by optimization or training generative neural networks.

BibTeX

@inproceedings{cvpr2021_aslicedwasserste,
  title = {A Sliced Wasserstein Loss for Neural Texture Synthesis},
  author = {Eric Heitz and Kenneth Vanhoey and Thomas Chambon and Laurent Belcour},
  booktitle = {CVPR 2021},
  year = {2021}
}