Anomaly detection methods based on GAN: a survey

Anomaly detection (AD) is an enduring topic, and it has been used in various fields, such as fraud detection, industrial fault diagnosis, and medical image diagnosis. With the continuous development of deep learning, in recent years, an increasing number of researchers have begun to use GAN-based methods to solve AD problems. In this article, we first classify these GAN-based anomaly detection (GBAD) methods according to the different forms of data. Afterward, we explain the advantages and limitations of each method in each category. Then, for each category, we quantitatively give the effects of these methods. Finally, we discuss the current open issues and possible research directions when using the GBAD method to solve anomaly detection problems in the future.