Document Type : Research Paper

Authors

• Rasha Muthana ¹
• Nora Ahmed Mohammed ²
• Amir Aboubakr Shaker Mohmoud ³

¹ Department of Roads and Transport Engineering, College of Engineering, University of Al-Qadisiyah, Al-Qadisiyah, Iraq

² Department of Electrical and Communication Engineering, College of Engineering, University of Al- Qadisiyah, Al-Qadisiyah, Iraq

³ School of Digital Forensics and Cyber Security, National Forensic Sciences University Gandhinagar, India

Abstract

Steganography is essential in modern cryptography and communications, enhancing the security and confidentiality of sensitive data exchanges. It has become an interesting tool because not only have security requirements for secret messages become stronger, but video has also become more popular. This paper introduces an advanced method combining sequence-to-sequence transformer models for speech recognition, RC4 encryption, and the Least Significant Bit (LSB) technique for data embedding in videos. The approach securely embeds audio messages within video streams, ensuring that even if detected, the data remains inaccessible without the decryption key. Our methodology includes converting audio to text, encrypting it, and embedding the encrypted data into video files, with a subsequent recovery process that preserves the original audio's emotional and tonal qualities. Evaluations using the UCF101 dataset confirm the method's effectiveness in maintaining video quality, with minimal visual distortion, and robust data security. This research provides a secure framework for covert communication, with potential applications in areas requiring high-level data privacy.

Keywords

• Steganography
• RC4 algorithm
• Least Significant Bit
• Forbidden Zone Data Hiding
• Human Auditory System