Enhancing Data Security in video Steganography using Face Recognition

In today's modern digitized world, the proliferation of communication technology and the internet has considerably increased the transmission of data in large amounts. Hence, secure data transmission is becoming a pre-requisite. Digital Steganography provides the capability to protect private communication in today's Internet era. Steganography is a practice to shelter and conceal multimedia information in a disguised manner, or we can say it is the study of invisible communication. Steganography uses image, text, video, and audio to disguise secret information. The presented system is an amalgamation of face recognition and LSB based steganography with spread spectrum technology. In the encryption stage, the extracting the audio from the original multimedia video file. Then secrete message is inserted and embedded in the extracted audio. The audio steganography uses the Least Significant Bit (LSB) substitution algorithm with a spread spectrum technique to add the secrete audio. The processed audio is called stego audio. The stego audio is embedded back into the video frames. Along with this, attach a face image of the authorized user is to the original video. The processed video is transmitted over the communication channel. In the decryption stage, the face image is extracted from the image and recognized by using the CNN algorithm. The audio can then be decrypted using a 4-bit LSB decryption algorithm only if the two previous facial images match. Thus the introduction of face recognition makes the system more robust. Testing of the sustainability of the system, the statistical analysis of the embedding algorithm has been performed and checked that it not only maintains the visual similarity in the stego file but also leaves the other statistics unchanged after embedding. The performance of the proposed algorithm is assessed by PSNR, SSIM, and RMSE metrics.