An image autonomous selection encryption algorithm based on the delay exponential logistic chaotic model

In recent decades, many excellent holistic image encryption schemes have been proposed. However, considering the practical application environment, only specific content in the image needs to be encrypted. Therefore, this paper proposes a chaos-based image encryption algorithm for the specific content of images. First, we propose a new chaotic map called Delay Exponential Logistic Chaotic Model. The simulation experiments show that the chaotic model has superior chaotic properties and produces complex pseudo-random sequences. Second, we set a sliding window on the image while segmenting the image inside the window using the DeepLab V3 semantic segmentation model trained on the cityscape dataset. Finally, the regions that exist to specific contents are encrypted using the proposed encryption algorithm. The simulation experiments show that our chaotic encryption algorithm has favorable encryption performance.