Open Access

ARTICLE

Multisecurity GAN-Steganography-Blockchain for IoT-Cloud Self-Service Banking

Mangala Natampalli^1,2,*, Kruthika Gottikere Channagangaiah³, Bodi Eswara Reddy¹, Rajkumar Buyya⁴, Venugopal Kupanna Rajuk³, Sundaraja Sitharama Iyengar⁵, Lalit Mohan Patnaik⁶

1 Computer Science and Engineering, Jawaharlal Nehru Technological University Anantapur, Anantapuramu, India
2 Centre for Development of Advanced Computing, Pashan, Pune, India
3 Computer Science and Engineering, University of Visvesvaraya College of Engineering, Bangalore University, Bangalore, India
4 Cloud Computing and Distributed Systems Laboratory, University of Melbourne, Melbourne, VIC, Australia
5 Computer Science, Florida International University, Miami, FL, USA
6 Consciousness Studies Program, National Institute of Advanced Studies, Bangalore, India

* Corresponding Author: Mangala Natampalli. Email:

Journal on Internet of Things 2026, 8, 1-30. https://doi.org/10.32604/jiot.2026.067726

Received 11 May 2025; Accepted 18 December 2025; Issue published 24 February 2026

Abstract

Contemporary banking focuses on self-service and customer-centric experience by harnessing Internet of Things (IoT) and Cloud Computing. However, these systems remain vulnerable to multifaceted cyberattacks. The IoT-Cloud-based systems can be safeguarded through authentication, confidentiality, integrity, availability, privacy, and non-repudiation. This work proposes a multi-stage security consisting of Generative Adversarial Networks (GAN) for facial authentication, hybrid Curvelet and Least-Significant-Bit (LSB) Steganography for data protection, and Ethereum Blockchain for transactions and storage security to provide complete protection to the self-service pipeline. Customers are authenticated using live images from IoT cameras by GAN facial recognition. Improved data concealment is provided by combining Curvelet and LSB Steganography. Ethereum Blockchain is selected for use in off-chain mode to provide integrity, time-stamping, non-repudiation, and fraud resistance for banking transactions, along with immutable permanent data storage in the Cloud. The Frechet Inception Distance and Inception Score metrics show that the GAN-generated front angle images are accurate, and the Precision, Recall, and F-Measure of 90% ascertain that the facial authentication is correct. Superior imperceptibility of the hybrid Curvelet-LSB Steganography is ascertained by the Peak-Signal-to-Noise-Ratio metric and Histogram analysis. An upload time of less than 3 minutes for fairly large data sizes, and cost-free transactions make Ethereum off-chain a preferred choice. Thus, the proposed multisecurity pipeline assures a highly secure and convenient banking experience for customers.

---

Keywords

---