Skip to content
Conference Open access

MultiLevelGAN: A Secure Adversarial Framework for Image Encryption and Protected Communications

Oct 2026 · International Conference on Innovative Mechanisms for Industry Applications · pp. 1722-1728 · 0 citations · 21 references

Abstract

With digital communication networks, cloud platforms and connected devices expanding rapidly, data security has never been as important. Cryptography techniques which for so long proved to be effective now face increasing opposition as attackers advance and computing power continues to accelerate in its abilities. Scientists are, therefore, now examining cryptography techniques which evolve over time and learn to outwit any such developments made by attackers. Generative Adversarial Network (GAN) is one such method. They work via an adversarial training approach to produce encryption patterns and keys that cannot be predicted, increasing security in a method that conventional cryptographic tools cannot manage. This article explains how the use of GANs can help transform image encryption and secure communication. We discuss working of generator and discriminator networks as models that offer to produce random and adaptive encryption. The impact beyond cryptography will also be discussed. How GAN-empowered methods can be utilized to make secure IoT devices and patient and healthcare data and secure future-generation networks is discussed here in the article. However, there are issues with GAN approaches of unstable GAN training; heavy computational requirement is the major limiting factor of GAN based encryption; scaling and adversarial attacks, but with proper planning these challenges can prove fruitful. Our results indicated that as opposed to conventional encryption techniques that do not deliver randomness, GAN based encryptions produces and withstands attacks better.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.