Algorithm-Dependent Computational Ghost Encryption and Imaging

Abstract

We demonstrate an algorithm-dependent computational ghost imaging (GI) by proposing a pattern multiplexing method. Based on the characteristics of three typical GI algorithms (i.e., traditional second-order GI, differential GI, total variational regularization GI), we design a set of pseudorandom patterns to achieve different reconstructed images with different GI algorithms. With our designed patterns, total variational regularization GI can be used for imaging unknown scenes or prehidden images, while traditional second-order GI and differential GI are used for encrypt other different images. Not only being critical for imaging quality in GI, reconstruction algorithms in our simulations and experiments demonstrate their potential role in multiple-image reconstruction, providing an idea for the application of GI in image encryption.