Collapse Resistant Deep Convolutional GAN for Multi-Object Image Generation

TL;DR

This paper presents a new conditional deep convolutional GAN architecture capable of generating multi-object images with diverse classes, realistic arrangements, and resistance to mode collapse, advancing multi-object image synthesis.

Contribution

The work introduces a novel, collapse-resistant conditional DCGAN architecture that learns to generate complex multi-object images with realistic spatial relationships.

Findings

Successfully generates multi-object images with diverse class combinations

Demonstrates stability against mode collapse during training

Produces images with realistic object arrangements and occlusions

Abstract

This work introduces a novel system for the generation of images that contain multiple classes of objects. Recent work in Generative Adversarial Networks have produced high quality images, but many focus on generating images of a single object or set of objects. Our system addresses the task of image generation conditioned on a list of desired classes to be included in a single image. This enables our system to generate images with any given combination of objects, all composed into a visually realistic natural image. The system learns the interrelationships of all classes represented in a dataset, and can generate diverse samples including a set of these classes. It displays the ability to arrange these objects together, accounting for occlusions and inter-object spatial relations that characterize complex natural images. To accomplish this, we introduce a novel architecture based on Conditional Deep Convolutional GANs that is stabilized against collapse relative to both mode and condition. The system learns to rectify mode collapse during training, self-correcting to avoid suboptimal generation modes.