Generation of multiple N shares based on deep learning protocol for efficient storage of patient health records data splitting in cloud framework

This work proposes a novel deep learning protocol for efficient storage of patient health records (PHRs) in a cloud framework. It mainly consists of uploading PHR to the cloud and retrieving PHR from the cloud. Before uploading the PHR, the text data encryption is done using Rivest, Shamir, and Adelman (RSA) algorithm. The PHR image encryption is carried out using the image encryption algorithm and the discrete cosine transform (DCT) is utilised for the compression. Then the PHR is shared to the cloud using the dense deep belief network (DenseBN)-based multiple 'N' sharing protocol. For the retrieval, the requested files are obtained by combining all partitioned parts of the shared data, followed by a decryption process. The DenseBN offers a key complexity of 0.765, key sharing time of 0.465, key retrieval time of 0.555, Jaccard similarity of 0.909, and memory of 0.374 for considering the key size as 64.