Author(s) :

K. Ranjini, E. Jayapriya

Conference Name :

International Conference on Modern Trends in Engineering and Management (ICMTEM-25)

Abstract :

Facial image steganography has emerged as a promising technique for safeguarding biometric data, particularly in printed ID systems vulnerable to photograph substitution attacks. This study proposes an end to end deep convolutional neural network (DCNN) framework for embedding steganographic facial data into printed ID photographs to enhance security and authenticity. The system leverages a dual stage architecture consisting of an encoder decoder pair, where the encoder embeds a concealed payload into the visible ID photograph, and the decoder retrieves this hidden information for verification. To ensure imperceptibility, the embedded data remains visually indistinguishable while being robust to common degradations like printing, scanning, and noise. This framework bridges the gap between biometric security and steganographic techniques, ensuring that any tampering with the ID photograph is detectable, thus preventing unauthorized identity manipulation. By combining advanced DCNNs with innovative steganographic principles, this research establishes a new benchmark in securing printed ID systems against photograph substitution attacks. This paper presents a Python Flask based application designed to detect and mitigate such attacks. By combining image analysis techniques, steganographic extraction, and machine learning, the application identifies hidden patterns that indicate suspicious activities. The proposed system offers a lightweight, scalable solution for real time detection, contributing to enhanced cybersecurity in authentication frameworks.

No. of Downloads :

1