Hyperledger blockchain enabled secure medical record management with deep learning-based diagnosis model

【Author】 Sammeta, Naresh; Parthiban, Latha

【Source】COMPLEX & INTELLIGENT SYSTEMS

【影响因子】6.700

【Abstract】In recent times, advanced developments in healthcare sector result in the generation of massive amounts of electronic health records (EHRs). EHR system enables the data owner to control his/her data and share it with designated people. The vast volume of data in the healthcare system makes it difficult for data to ensure security and diagnostic processes. To resolve these issues, this paper develops a new hyperledger blockchain enabled secure medical data management with deep learning (DL)-based diagnosis (HBESDM-DLD) model. The presented model involves distinct stages of operations such as encryption, optimal key generation, hyperledger blockchain-based secure data management, and diagnosis. The presented model allows the user to control access to data, permit the hospital authorities to read/write data, and alert emergency contacts. For encryption, SIMON block cipher technique is applied. At the same time, to improve the efficiency of the SIMON technique, a group teaching optimization algorithm (GTOA) is applied for the optimal key generation of the SIMON technique. Moreover, the sharing of medical data takes place using multi-channel hyperledger blockchain that utilizes a blockchain for storing patient visit data and for the medical institutions to record links for the EHRs saved in external databases. Once the data are decrypted at the receiving end, finally, variational autoencoder (VAE)-based diagnostic model is applied to detect the existence of the diseases. The performance validation of the HBESDM-DLD model takes place on benchmark medical dataset and the results are inspected under various performance measures. The experimental results proves that the HBESDM-DLD methodology is superior to state-of-the-art methods.

【Keywords】Blockchain; Electronic health records; SIMON; Variational autoencoder; Deep learning

【发表时间】

【收录时间】2022-01-01

【文献类型】

【主题类别】

--