查看文章

In recent days, the application of generative adversarial networks (GAN) shows a drastic growth in the various areas of research. The application of GAN in several areas delivers a noteworthy improvement on the performance analysis. The traditional methods have several limitations in solving the problems of pose variation. Fortunately, deep learning paradigm opens new ways of the person re-identification research based on pose variations and has become a hot topic in this field. In this work, we examine the general approaches of pose-guided image generation with the GAN. We also discuss different datasets and the performance evaluation of various approaches on GAN. We review the current person re-identification models using GAN and identify various limitations of these models. In this work, we provide a comprehensive analysis of various GAN approaches to give reader an overview of the latest research work in this area. We also identify some research gaps which will provide future research directions for researchers in this field.