Secure Framework for Patient Data Transmission on Mobile-Cloud Platform

Abstract

In this paper we have addressed the authentication as well as payload problem of medical image that is sent to the Cloud by resource constrained devices (like mobile phones) for selective encryption. Firstly, the given medical image is segmented into two distinct regions based on diagnostic importance i.e., non region of importance (NRoI) and region of importance (RoI) using Otsu thresholding technique. RoI obtained is then divided into four blocks and among those four blocks any two blocks are randomly embedded into two separate cover-images. In order to ensure authenticity of image carrying selective-data to be encrypted, fragile watermark is inserted in same cover-image. On reducing the selective-data payload will decrease which will effectively decrease time taken by Cloud for encryption (as Cloud has to encrypt RoI block only rather than encrypting the whole stego-image) which is a must criterion for real time applications. The Cloud looks out for embedded watermark once it receives stego-image from client to ensure authenticity of the stego-image. Embedded data in stego-image is extracted by Cloud and performs the encryption of RoI block only. Selectively encrypted image is sent back to client. Client then extracts encrypted RoI blocks from encrypted stego-images and combines the four RoI blocks (two encrypted and two unencrypted) to obtain the encrypted RoI. Finally encrypted RoI and NRoI are combined to obtain full medical image. The final medical image is then forwarded to medical centers, experts etc. for analysis plus diagnosis purposes. The proposed framework reduces payload which in turn reduces the size of image which needs encryption, thus saving Cloud resources. Authenticity of stego-image that is sent to Cloud is also ensured.