GAN-based models and applications
Jiaqi Li, Tujie Li, Ruoxuan Lin, Qizhou Nie · 2022 IEEE 5th International Conference on Information Systems and Computer Aided Education (ICISCAE) · 2022
GANs have achieved great success in image generation. Gan consists of two main parts, the generator and the discriminator, the generator tries to generate real samples that fool the discriminator, and the discriminator tries to distinguish between real samples and generated samples. Here The continuous improvement of simulation ability and generation ability under this kind of confrontation game can realize the generation of fake images. Since GAN appeared in 2014, articles of various types of GAN have been published in major journals and conferences, and the application of GAN in image generation (specifying image synthesis, text-to-image, image-to-image, video) and GAN in NLP and other areas of application are the most studied, and research in this area has demonstrated the great potential of using GANs in image synthesis. In this paper, the classical basic GAN model is introduced at first. Then, the paper analysis the differences and characteristics of the recent GAN-based models, and introduces the applications of GAN-based models in different tasks.