Deep Correlations for Texture Synthesis

Omry Sendik, Daniel Cohen-OrACM Trans. Graphics 2017 · DOI · OpenReview

Abstract

Example-based texture synthesis has been an active research problem for over two decades. Still, synthesizing textures with nonlocal structures remains a challenge. In this article, we present a texture synthesis technique that builds upon convolutional neural networks and extracted statistics of pretrained deep features. We introduce a structural energy, based on correlations among deep features, which capture the self-similarities and regularities characterizing the texture. Specifically, we show that our technique can synthesize textures that have structures of various scales, local and nonlocal, and the combination of the two.
BibTeX
@article{sendik2017deep,
  author = {Omry Sendik and Daniel Cohen-Or},
  title = {{Deep Correlations for Texture Synthesis}},
  journal = {ACM Trans. Graphics},
  year = {2017},
  doi = {10.1145/3015461},
}

Loading related content...

Figure