Image encryption algorithm based on face recognition, facial features recognition and bitonic sequence

Image encryption algorithm based on face recognition, facial features recognition and bitonic sequence

연구 분야: Artificial Intelligence

학회: Multimedia Tools and Applications

Traditional scrambling algorithms frequently rely on static and fixed scrambling modes, which lack the involvement of chaotic sequences during the scrambling phase. This results in poor randomness in the scrambling process and can leave key information, such as facial features in images, inadequately protected. In the event that such sensitive information is stolen, it could lead to significant trouble. To mitigate these issues, this paper presents an image encryption algorithm that incorporates face recognition and bitonic sequence techniques. The algorithm utilizes the SHA-512 (Secure Hash Algorithm) for key generation and the Chen system for generating chaotic sequences during the encryption process. Initially, the algorithm identifies the face and facial features within the image via face recognition and facial feature recognition technologies. A row-column scrambling algorithm, designed based on the characteristics of the bitonic sequence, is then implemented to scramble the facial features while the Zigzag algorithm is used to break the row-column correlation. With respect to the overall image scrambling, the Fisher Yeats scrambling algorithm is employed, and the entire image is uniformly diffused. Through simulation experiments and security tests, the proposed algorithm has shown better performance than other methods in terms of NPCR and UACI testing studies, resulting in outcomes closer to the ideal values of 99.6094% and 33.4635%, respectively. Other experimental data also demonstrates performance that is near ideal, and the decrypted images show good visual quality against various attacks. Overall, the proposed algorithm exhibits strong robustness.