Aligning Latent and Image Spaces to Connect the Unconnectable
Authors:
- 1 October 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 14124-14133
Abstract
In this work, we develop a method to generate infinite high-resolution images with diverse and complex content. It is based on a perfectly equivariant patch-wise generator with synchronous interpolations in the image and latent spaces. Latent codes, when sampled, are positioned on the coordinate grid, and each pixel is computed from an interpolation of the neighboring codes. We modify the AdaIN mechanism to work in such a setup and train a GAN model to generate images positioned between any two latent vectors. At test time, this allows for generating infinitely large images of diverse scenes that transition naturally from one into another. Apart from that, we introduce LHQ: a new dataset of 90k high-resolution nature landscapes. We test the approach on LHQ, LSUN Tower and LSUN Bridge and outperform the baselines by at least 4 times in terms of quality and diversity of the produced infinite images. The project website is located at https://universome.github.io/alis.Keywords
All Related Versions
This publication has 18 references indexed in Scilit:
- TileGANACM Transactions on Graphics, 2019
- Wide-Context Semantic Image Extrapolation2019
- A Style-Based Generator Architecture for Generative Adversarial Networks2019
- Arbitrary Style Transfer in Real-Time with Adaptive Instance Normalization2017
- Improved Texture Networks: Maximizing Quality and Diversity in Feed-Forward Stylization and Texture Synthesis2017
- Image Style Transfer Using Convolutional Neural Networks2016
- Infinite Images: Creating and Exploring a Large Photorealistic Virtual SpaceProceedings of the IEEE, 2010
- Image Alignment and Stitching: A TutorialFoundations and Trends in Computer Graphics and Vision, 2007
- Procedural modeling of buildings2006
- Procedural modeling of cities2001