Verifiable Attribute-Based Through Text and Image Search with Fine-Grained Owner-Enforced Search Authorization in the Cloud

Search over encrypted data is a critically important enabling technique in cloud computing, where encryption-before outsourcing is a crucial answer for ensuring user information security in the untrusted cloud server condition. In this work we focus on Role- based authentication is a combination of symmetric key cryptography and public key cryptography where by every encryption process needs Data, Public Key, Group Key, The Policy is a set of rule that can be specified as chain also for e. g. , in the context of hospital a policy can be given as {doctor, patient} which means anybody from the doctors group or patient group can simultaneously decrypt the document on the other hand a policy can be specified as chain policy for instance{doctor}, {patient} This is known as nested policy. In such cases a patient can decrypt the document only when that is decrypted by doctors first, Policy based encryption is becoming popular in an enterprise context where different authorities require different permission and privacy settings for the access of the records. In this work we have develop a novel CP-ABE (Cryptography-Attributed based Encryption) based technique in an enterprise hospital context deployed in a local cloud with fog computing architecture. Our proposed system provides different level of encryption, decryption, authentication, authorization and privacy setting for doctors in the context of patients, medical image, image feature records. The data used for communication between doctor and patient is images of skin cancer and related symptoms. Both text and image is encrypted and decrypted by AES (Advanced Encryption Standard) during data transfer. Further machine learning Bayesian Classifier approach is used to diagnose abnormality in the skin cancer image. In this work we demonstrate the use of CP-ABE with human entities as well as in the context of machine learning. The overall work demonstrate the entire process of medical scanning in an enterprise hospital application both plain record encryption and image encryption with the help of CP-ABE. In order to demonstrate the efficiency of the system we have developed a simple medical application where doctor encrypt text and image which can visualize only their own patients data and the usability of machine learning image classification technique which automatically can retrieve the observation of the image as abnormal or normal. This proves machine learning and performance evaluation shows efficiency of our system.