An Automated Group Key Authentication System Using Secret Image Sharing Scheme

TL;DR

This paper presents an automated group key authentication system that uses secret image sharing and a classifier to ensure secure group communication in open networks, achieving high accuracy in validating reconstructed key images.

Contribution

The work introduces a novel authentication system combining secret image sharing with a classifier for automatic validation of group keys in secure communication.

Findings

Achieved 99.1% classification accuracy for characters and digits.

Developed a system for secure group key validation using image sharing and machine learning.

Validated the system's effectiveness in open network environments.

Abstract

In an open network environment, privacy of group communication and integrity of the communication data are the two major issues related to secured information exchange. The required level of security may be achieved by authenticating a group key in the communication channel, where contribution from each group member becomes a part of the overall group key. In the current work, we have developed an authentication system through Central Administrative Server (CAS) for automatic integration and validation of the group key. For secured group communication, the CAS generates a secret alphanumeric group key image. Using secret image sharing scheme, this group key image shares are distributed among all the participating group members in the open network. Some or all the secret shares may be merged to reconstruct the group key image at CAS. A k-nearest neighbor classifier with 48 features to represent the images, is used to validate the reconstructed image with the one stored in the CAS. 48 topological features are used to represent the reconstructed group key image. We have achieved 99.1% classification accuracy for 26 printed English uppercase characters and 10 numeric digits.