Megapixel Size Image Creation using Generative Adversarial Networks

Abstract

Since its appearance, Generative Adversarial Networks (GANs) have received a lot of interest in the AI community. In image generation several projects showed how GANs are able to generate photorealistic images but the results so far did not look adequate for the quality standard of visual media production industry. We present an optimized image generation process based on a Deep Convolutional Generative Adversarial Networks (DCGANs), in order to create photorealistic high-resolution images (up to 1024x1024 pixels). Furthermore, the system was fed with a limited dataset of images, less than two thousand images. All these results give more clue about future exploitation of GANs in Computer Graphics and Visual Effects.

3 pages, 4 figures

References in corpus (3)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• Improved Techniques for Training GANs

Cited by in corpus (12)

• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Training Generative Adversarial Networks with Limited Data
• Improved Precision and Recall Metric for Assessing Generative Models
• TDEFSI: Theory Guided Deep Learning Based Epidemic Forecasting with Synthetic Information
• Truncated Log-concave Sampling with Reflective Hamiltonian Monte Carlo
• SeismoGen: Seismic Waveform Synthesis Using Generative Adversarial Networks
• Ebola Optimization Search Algorithm (EOSA): A new metaheuristic algorithm based on the propagation model of Ebola virus disease
• Instance Selection for GANs
• Neural Crossbreed: Neural Based Image Metamorphosis
• Conditioning Trick for Training Stable GANs
• Exploiting Spatial Dimensions of Latent in GAN for Real-time Image Editing
• A Customizable Dynamic Scenario Modeling and Data Generation Platform for Autonomous Driving