Secure Image Cloud Storage Using Homomorphic Password Authentication with ECC Based Cryptosystem

Abstract

Cloud based storage gives us ease of storing large data. Cloud is a storage that can store multimedia images and other data, but what happens to the data after the storage is invisible to the user. To protect the image in the cloud and transmission to the cloud, image encryption is used. Confidentiality is the way to protect from unauthorized access while transmitting and storing medical images, military images, online banking and telecommunication. Recently, chaos has become an emerging trend towards the encryption of images. New techniques are adopted in image encryption to increase the randomness and to decrease the correlation between the pixels. In the present work, an authentication based framework and image cryptosystem has been proposed for secure storage in the cloud. Authentication is provided using SHA hash algorithm and elgamal encryption. Image cryptosystem uses chaos based scrambling using bit-matrix generated from logistic chaotic map sequence. A novel key generation for chaos is done using Fourier transform for the image cryptosystem. Bit sequence and sorting based pixel permutation are performed using the chaotic sequence. Totalistic cellular automata technique has been applied using the corner neighborhood values to diffuse the image pixels. Pixels of the image is finally diffused using the random sequence formed using elliptic curve points/ coordinates to improve the security. This encryption technique is evaluated using statistical and differential analysis. ECC gives security using the discrete logarithm problem. From the analyses, it is shown that it possesses large keyspace, sensitivity towards a small change in pixel value and also key value. It also resists pattern attacks as well as differential attacks.