Generating Images Part by Part with Composite Generative Adversarial Networks

Abstract

Image generation remains a fundamental problem in artificial intelligence in general and deep learning in specific. The generative adversarial network (GAN) was successful in generating high quality samples of natural images. We propose a model called composite generative adversarial network, that reveals the complex structure of images with multiple generators in which each generator generates some part of the image. Those parts are combined by alpha blending process to create a new single image. It can generate, for example, background and face sequentially with two generators, after training on face dataset. Training was done in an unsupervised way without any labels about what each generator should generate. We found possibilities of learning the structure by using this generative model empirically.

References in corpus (4)

• Auto-Encoding Variational Bayes
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• Generating Images from Captions with Attention

Cited by in corpus (12)

• RSGAN: Face Swapping and Editing using Face and Hair Representation in Latent Spaces
• BlockGAN: Learning 3D Object-aware Scene Representations from Unlabelled Images
• Transformation-Grounded Image Generation Network for Novel 3D View Synthesis
• Investigating Object Compositionality in Generative Adversarial Networks
• Emergence of Object Segmentation in Perturbed Generative Models
• Attention-GAN for Object Transfiguration in Wild Images
• FineGAN: Unsupervised Hierarchical Disentanglement for Fine-Grained Object Generation and Discovery
• MOGAN: Morphologic-structure-aware Generative Learning from a Single Image
• Depth Structure Preserving Scene Image Generation
• Illiterate DALL-E Learns to Compose
• Generating Object Stamps
• An Interpretable Generative Model for Handwritten Digit Image Synthesis