Generative Adversarial Networks and Other Generative Models

TL;DR

This paper provides an overview of Generative Adversarial Networks (GANs), their development, training challenges, and applications in image generation and analysis, contrasting them with other generative models and discussing their limitations and advantages.

Contribution

It offers a comprehensive introduction to GANs, including their motivation, training stability techniques, and applications, along with a comparison to other generative models.

Findings

GANs have been successful in producing realistic images.

Stable training methods are crucial for effective GAN performance.

GANs can be applied to image segmentation and anomaly detection.

Abstract

Generative networks are fundamentally different in their aim and methods compared to CNNs for classification, segmentation, or object detection. They have initially not been meant to be an image analysis tool, but to produce naturally looking images. The adversarial training paradigm has been proposed to stabilize generative methods, and has proven to be highly successful -- though by no means from the first attempt. This chapter gives a basic introduction into the motivation for Generative Adversarial Networks (GANs) and traces the path of their success by abstracting the basic task and working mechanism, and deriving the difficulty of early practical approaches. Methods for a more stable training will be shown, and also typical signs for poor convergence and their reasons. Though this chapter focuses on GANs that are meant for image generation and image analysis, the adversarial training paradigm itself is not specific to images, and also generalizes to tasks in image analysis. Examples of architectures for image semantic segmentation and abnormality detection will be acclaimed, before contrasting GANs with further generative modeling approaches lately entering the scene. This will allow a contextualized view on the limits but also benefits of GANs.