An efficient algorithm to generate dynamic substitution-boxes and its applications in image encryption

Nowadays, securing private information against modern attacks is a challenging task. The construction of substitution-box (S-box) generators and encryption schemes, which meet security requirements against all kinds of attacks, has become a hot research topic in recent decades. Various algorithms are designed to generate S-boxes to increase the security of block ciphers. However, many of the existing S-box generators have shortcomings such as low nonlinearity values and high computation time. This research proposes a new elliptic curve (EC) and Möbius transformation based S-box generator and encryption scheme. Our method generates computationally efficient and highly dynamic S-boxes with good cryptographic properties, and the encryption scheme is robust against modern cryptographic attacks. Our method generates S-boxes with a nonlinearity value of 112 and takes only 0.367996 s to generate an S-box. Further, the average nonlinearity of a set of randomly-generated 10,000 of our S-boxes is 103.53, much better than state-of-the-art generators. The proposed encryption scheme is rigorously validated through comprehensive testing: NIST suite evaluations, key sensitivity, correlation analysis, and vulnerability to various attacks, yielding notable results: entropy value of 7.9977, NPCR of 99.54, and UACI of 33.69 for the encrypted Lena image. Thus, our proposed method exhibits high security against cryptographic attacks.