Low-Cost and Confidentiality-Preserving Multi-Image Compressed Acquisition and Separate Reconstruction for Internet of Multimedia Things

Abstract

Internet of Multimedia Things (IoMT) is facing how to achieve low-cost compression and acquisition of multimedia big data at the resource-constrained side while preserving the confidentiality of the data. In this article, we propose a low-cost and confidentiality-preserving multi-image compressed acquisition model and provide separate image reconstruction. We group a series of image sets and randomly sample them with compressive sensing in each group. It is noteworthy that we harness the sigmoid sequence to fuse the measurement of each group to alleviate the uneven distribution of the measurement. The experiment verifies that the proposed method avoids the energy information leakage of the original signal from the measurement. Subsequently, we assemble the sampling result of each group into a big master image with a suitable size and further encrypt it. The encrypted master image is transmitted to the cloud for storage and decryption service. The cloud performs a separate reconstruction of images. It provides different image reconstruction services for a fused measurement according to realistic demands, which reduces unnecessary resource consumption. Simulation results show the effectiveness and security of our acquisition scheme, which indicates the proposal can have a potential application to IoMT.