Secure Reconstruction of Image from Compressive Sensing in Cloud

Application of Large-scale images has been increased rapidly. Today the trend to use the cloud for storing data and take advantage of its computational capability for running complex algorithms have been grown. In outsourcing images two important issues must be considered: first, compression for efficient use of cloud’s storage that in our research compressive sensing (CS) used as secure method for reducing the capacity of image; second, be confident of privacy-assured by cloud service, because cloud as third party in communication between sender and user, should not be permitted to access the plain image that in many cases has privacy information. To assure this privacy, we propose a very easy and effective method that uses the sparseness of image and just by simple random permutation key, mapped the plain sparse vector into another sparse vector by applying permutation and substitution on components of plain sparse vector simultaneously. Afterward, this encrypted sparse vector is input of CS, and then the compressed encrypted image transmitted to cloud via unsecure channel. On-demand of user, cloud will reconstruct the compressed image and shares the encrypted image, real user by privacy key, could be able to decrypt it, even by very weak devices.

Keywords:

Sparse encryption , compressive sensing , image reconstruction