Publication year: 2023
ISBN: CCE00090
Internet Resource: Please Login to download book
This project introduces a chat application that utilizes deep learning-based steganography, providing an efficient and highly secure means of exchanging multimedia data. The chat application is designed to work on multiple platforms and provides a user-friendly interface, making it accessible to a wide range of users. The proposed solution employs steganography techniques to hide audio, image, and text data within the same medium. For audio steganography, we used an Encoder-Decoder paradigm trained on TIMIT dataset. We propose using the magnitude spectrogram instead of the power spectrogram that was used in all previous works, resulting in an improved secret audio signal-tonoise ratio (SNR) compared to previous works.
Subject: Computer sceince, Steganography, Deep learning, Intelligence security, End-to-end encryption