Non-Stationary Texture Synthesis by Adversarial Expansion

Abstract

The real world exhibits an abundance of non-stationary textures. Examples include textures with large-scale structures, as well as spatially variant and inhomogeneous textures. While existing example-based texture synthesis methods can cope well with stationary textures, non-stationary textures still pose a considerable challenge, which remains unresolved. In this paper, we propose a new approach for example-based non-stationary texture synthesis. Our approach uses a generative adversarial network (GAN), trained to double the spatial extent of texture blocks extracted from a specific texture exemplar. Once trained, the fully convolutional generator is able to expand the size of the entire exemplar, as well as of any of its sub-blocks. We demonstrate that this conceptually simple approach is highly effective for capturing large-scale structures, as well as other non-stationary attributes of the input exemplar. As a result, it can cope with challenging textures, which, to our knowledge, no other existing method can handle.

Accepted to SIGGRAPH 2018

References in corpus (2)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

Cited by in corpus (46)

• GAN-Leaks: A Taxonomy of Membership Inference Attacks against Generative Models
• Point2Mesh: A Self-Prior for Deformable Meshes
• Swapping Autoencoder for Deep Image Manipulation
• InSituNet: Deep Image Synthesis for Parameter Space Exploration of Ensemble Simulations
• TileGAN: Synthesis of Large-Scale Non-Homogeneous Textures
• Deep Learning-based Forgery Attack on Document Images
• Deep Geometric Texture Synthesis
• An Inverse Procedural Modeling Pipeline for SVBRDF Maps
• Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints
• Image Fine-grained Inpainting
• InGAN: Capturing and Remapping the "DNA" of a Natural Image
• SeamlessGAN: Self-Supervised Synthesis of Tileable Texture Maps
• On Demand Solid Texture Synthesis Using Deep 3D Networks
• Learning to Generate 3D Shapes from a Single Example
• Structural-analogy from a Single Image Pair
• TexSliders: Diffusion-Based Texture Editing in CLIP Space
• IEGAN: Multi-purpose Perceptual Quality Image Enhancement Using Generative Adversarial Network
• Texture Synthesis Guided Deep Hashing for Texture Image Retrieval
• ImageEye: Batch Image Processing Using Program Synthesis
• Texture Mixer: A Network for Controllable Synthesis and Interpolation of Texture
• TM-NET: Deep Generative Networks for Textured Meshes
• Controllable and Progressive Image Extrapolation
• Diverse Single Image Generation with Controllable Global Structure
• Texture Generation with Neural Cellular Automata
• Compositional Neural Textures
• Multi-Plane Program Induction with 3D Box Priors
• Data-Driven Modeling of Noise Time Series with Convolutional Generative Adversarial Networks
• TuiGAN: Learning Versatile Image-to-Image Translation with Two Unpaired Images
• Semantically Multi-modal Image Synthesis
• A Neural-Symbolic Architecture for Inverse Graphics Improved by Lifelong Meta-Learning
• Deep 3D-to-2D Watermarking: Embedding Messages in 3D Meshes and Extracting Them from 2D Renderings
• Conditional Entropy Coding for Efficient Video Compression
• Patternshop: Editing Point Patterns by Image Manipulation
• Neural Photometry-guided Visual Attribute Transfer
• Level generation and style enhancement -- deep learning for game development overview
• Content-aware Tile Generation using Exterior Boundary Inpainting
• DecoSurf: Recursive Geodesic Patterns on Triangle Meshes
• Towards Controllable and Photorealistic Region-wise Image Manipulation
• DiffFit: Visually-Guided Differentiable Fitting of Molecule Structures to a Cryo-EM Map
• Directional Texture Editing for 3D Models
• A Neural-Symbolic Framework for Mental Simulation
• TE141K: Artistic Text Benchmark for Text Effect Transfer
• Geometry Enhancements from Visual Content: Going Beyond Ground Truth
• Seamless Satellite-image Synthesis
• Projective Urban Texturing
• Deep Tiling: Texture Tile Synthesis Using a Deep Learning Approach