Incorporating long-range consistency in CNN-based texture generation

2016/06/03 by Berger, G., Memisevic, R. · 1 citation
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.1606.01286

Abstract

Gatys et al. (2015) showed that pair-wise products of features in a convolutional network are a very effective representation of image textures. We propose a simple modification to that representation which makes it possible to incorporate long-range structure into image generation, and to render images that satisfy various symmetry constraints. We show how this can greatly improve rendering of regular textures and of images that contain other kinds of symmetric structure. We also present applications to inpainting and season transfer.

Cited by

• Neural Artistic Style and Color Transfer Using Deep Learning

Related

• High resolution neural texture synthesis with long range constraints
• Texture Synthesis Through Convolutional Neural Networks and Spectrum Constraints
• Texture Synthesis Using Shallow Convolutional Networks with Random Filters
• Texture Networks: Feed-forward Synthesis of Textures and Stylized Images
• Texture Synthesis Using Convolutional Neural Networks