Imperceptible and Robust Encryption: Salsa20 Stream Cipher for Colour Image Data

Abstract

Data security has become crucial, especially in today's era, therefore we need to protect our personal data to avoid unwanted incidents. The primary objective of this research is to empirically demonstrate the viability of our proposed methodology for encrypting color images using the Salsa20 algorithm, renowned for its stream cipher characteristics, which inherently afford it a swift processing speed. The encryption method we use is to take each pixel from the original image and convert it into bytes based on the RGB value in it, then encrypt it using a keyword that has been converted using a hash function. In this study, we carried out several evaluations to evaluate the performance of the encrypted and decrypted images to test the method we propose, including histogram analysis and compare patterns, visual image testing, and key space analysis. Through this experiment, it has been proven that Salsa20 is effective in maintaining confidentiality and image integrity. Histogram analysis reveals differences in pixel distribution patterns between the original and encrypted images. Visual testing shows that the encrypted image maintains good optical quality. Keyspace analysis ensures the security of encryption keys. The performance evaluation resulted in an NPCR above 99%, UACI had been reached 69.28%, MSE was closes to 0, and the highest PSNR was around 61.89dB, this shows that encrypted images recovered with high accuracy.