Deep Correlations for Texture Synthesis
Authors:
Open Access
- 25 July 2017
- journal article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Graphics
- Vol. 36 (4) , 161
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.Keywords
This publication has 23 references indexed in Scilit:
- Image meldingACM Transactions on Graphics, 2012
- Image meldingACM Transactions on Graphics, 2012
- Synthesizing structured image hybridsACM Transactions on Graphics, 2010
- Synthesizing structured image hybridsACM Transactions on Graphics, 2010
- Appearance-space texture synthesisACM Transactions on Graphics, 2006
- Texture optimization for example-based synthesisACM Transactions on Graphics, 2005
- Parallel controllable texture synthesisACM Transactions on Graphics, 2005
- Near-regular texture analysis and manipulationACM Transactions on Graphics, 2004
- Reduced multidimensional co-occurrence histograms in texture classificationIEEE Transactions on Pattern Analysis and Machine Intelligence, 1998
- A Limited Memory Algorithm for Bound Constrained OptimizationSIAM Journal on Scientific Computing, 1995