Research Developments in Generative Adversarial Networks for Image Restoration and Communication
Generative Adversarial Networks (GAN) is always a popular study topic in artificial intelligence. This paper will analyze the principle of the GAN and introduce the development of the GAN and various derivative models. The improved Super-Resolution GAN (SRGAN) model and Cascading Residual Super-Reso...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
EDP Sciences
2025-01-01
|
| Series: | ITM Web of Conferences |
| Online Access: | https://www.itm-conferences.org/articles/itmconf/pdf/2025/01/itmconf_dai2024_01020.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Generative Adversarial Networks (GAN) is always a popular study topic in artificial intelligence. This paper will analyze the principle of the GAN and introduce the development of the GAN and various derivative models. The improved Super-Resolution GAN (SRGAN) model and Cascading Residual Super-Resolution GAN (CR-SRGAN) model based on the GAN model achieve super-resolution of dark and old artifact images and solve the problem of color restoration and texture enrichment of dark and old artifacts. The GAN model is also widely used in the field of communication and information security. It proposes an End-to-End(E-to-E) communication encryption system based on Deep Convolutional GAN (DCGAN) to solve the secure transmission problem in wireless communication systems based on E-to-E learning. The system can realize encoding and decoding of input bits of arbitrary length with good generalization ability. Finally, the image restoration and communication encryption are summarized, along with an outlook on their development trends. |
|---|---|
| ISSN: | 2271-2097 |