Technical gazette, Vol. 31 No. 6, 2024.
Original scientific paper
https://doi.org/10.17559/TV-20240304001373
A Blockchain and Hybrid Deep Learning for Secure and Efficient Healthcare Data Transmission and Management
K. Kiruthikadevi
; Department of Computer Science and Engineering, Nandha College of Technology, Erode - 638052
*
R. Sivaraj
; Department of Computer Science and Engineering, Nandha College of Technology, Erode - 638052
M. Vijayakumar
; Department of Computer Science and Engineering, Sasurie College of Engineering, Vijayamangalam - 638056
* Corresponding author.
Abstract
The tremendous growth in technology has laid the foundations for more efficient solutions in the healthcare field, aiming to optimize security and scalability while improving patient care. This abstract presents an advanced methodology, leveraging hybrid deep learning techniques and blockchain (BC) technology to revolutionize healthcare system. BC technology provides a transparent and decentralized framework, which allows secured data sharing, storage, and access control. By incorporating BC into healthcare systems, interoperability, data integrity, and privacy can be ensured while disregarding the reliance on central authority. In combination with BC, hybrid DL methods provide powerful abilities for decision making and data analysis in healthcare. Integrating the strengths of deep learning (DL) techniques with classical machine learning methodologies, hybrid DL allows efficient and accurate processing of complicated healthcare data, such as sensor data, medical records, and images. This study presents a Blockchain with Deep Learning Assisted Data Transmission and Classification (BDC-DTC) technique in the healthcare sector. The presented BDC-DTC technique involves the design of image encryption with BC technology for achieving security in the healthcare sector. Initially, Elgamal encryption approach is used to encrypt the medical images which are then stored securely using BC technology. Next, the disease detection process is carried out using multi-faceted approach namely residual network (ResNet18) feature extractor, weIghted meaNoFvectOrs (INFO) based hyperparameter selection, and backpropagation neural network (BPNN) based classification. The simulation results of the BDC-DTC method can be studied using medical image database. The experimental outcomes specified that the BDC-DTC method gains superior performance over other models in terms of distinct measures.
Keywords
blockchain technology; deep learning; disease detection; healthcare data security; medical image encryption
Hrčak ID:
321952
URI
Publication date:
31.10.2024.
Visits: 6 *