Texture Mixer: A Network for Controllable Synthesis and Interpolation of Texture
Preprint
Abstract
This paper addresses the problem of interpolating visual textures. We formulate this problem by requiring (1) by-example controllability and (2) realistic and smooth interpolation among an arbitrary number of texture samples. To solve it we propose a neural network trained simultaneously on a reconstruction task and a generation task, which can project texture examples onto a latent space where they can be linearly interpolated and projected back onto the image domain, thus ensuring both intuitive control and realistic results. We show our method outperforms a number of baselines according to a comprehensive suite of metrics as well as a user study. We further show several applications based on our technique, which include texture brush, texture dissolve, and animal hybridization.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- Non-stationary texture synthesis by adversarial expansionACM Transactions on Graphics, 2018
- The Unreasonable Effectiveness of Deep Features as a Perceptual Metric2018
- Generative Image Inpainting with Contextual Attention2018
- Arbitrary Style Transfer in Real-Time with Adaptive Instance Normalization2017
- Deep Correlations for Texture SynthesisACM Transactions on Graphics, 2017
- Inception-v4, Inception-ResNet and the Impact of Residual Connections on LearningProceedings of the AAAI Conference on Artificial Intelligence, 2017
- Synthesis of Complex Image Appearance from Limited ExemplarsACM Transactions on Graphics, 2015
- Space-Time Completion of VideoIEEE Transactions on Pattern Analysis and Machine Intelligence, 2007
- Texture synthesis by non-parametric sampling1999