Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a novel image encryption algorithm that leverages Convolutional Neural Networks (CNNs) to dynamically generate adaptive S-boxes, enhancing security against linear and differential attacks. By training the CNN to learn characteristics of input images, the proposed method achieves greater non-linearity and uniqueness in S-box generation compared to conventional fixed S-boxes. Security assessments using entropy, histogram analysis, and other metrics demonstrate that this dynamic approach significantly improves resilience and flexibility in image encryption.
CNN-generated S-boxes offer a breakthrough in image encryption, providing enhanced security and adaptability that outperforms traditional methods.
The paper proposes a dynamic approach to image encryption, combining the use of Convolutional Neural Networks (CNNs) and classical cryptography to improve the security and flexibility of image encryption. The main concept is to create adaptive Substitution boxes (S-boxes) based on characteristics that are learned by a trained CNN. The CNN-based S-boxes can be relied on for more non-linearity, uniqueness, and input image dependence than the conventional fixed S-boxes because they are susceptible to the linear and differential attacks. This dynamic behaviour enhances the confusion property and makes it more resistant to statistical and structural attacks. The encryption algorithm consists of CNN-based feature extraction and the creation of a personalised S-box to replace the pixels. Entropy, histogram analysis, correlation, NPCR, and UACI enable security assessment of generated S-boxes based on the CNN, indicating that the scheme is more resilient and flexible than traditional ones.