Comparative Assessment of Hash Functions in Securing Encrypted Images

Abstract

Different encryption methods have been developed to securely transmit confidential images over the Internet and combat the increasing cybercrime. Many of these methods use hash functions to enhance encryption strength. Due to the lack of a comprehensive evaluation of how different hash functions affect image encryption, this study presents a comparative analysis of the performance of various hash functions as encryption keys and analyzes their security, speed, and efficiency. The source image is first processed as a series of bytes. The bytes are divided into byte vectors, each with a length that matches the length of the hash value of a specified hash function. An XOR operation is performed between the hash value bytes and the associated byte vector. The bytes are reordered in each vector according to the ascending order of the associated hash value. Several metrics, such as Normalized Mean Absolute Error (NMAE), Peak Signal to Noise Ratio (PSNR), entropy, key size, and hash time, were used to evaluate the performance of different hash functions in image encryption. The results showed a clear variation in using various hash functions in terms of security, speed, and efficiency. With NMAE>72%, PSNR<6.62 dB, and Entropy>7.999 bpp, the use of the SHA family and MD5 is recommended in applications that need to achieve a high level of distortion in encrypted images. To resist brute-force attacks on the key, Blake2b, SHA512, and Whirlpool are the best choices with a key size of 512 bits. The Tiger is the fastest hash function, requiring the least average time of 0.372 seconds to complete the encryption process, making it the best choice for real-time applications. These findings help to choose the appropriate hash function in developing cryptographic techniques for a particular area.