Peyvandi Amirhossein, Majidi Babak, Peyvandi Soodeh, Patra Jagdish
Department of Computer Engineering, Khatam University, Tehran, Iran.
Emergency and Rapid Response Simulation (ADERSIM) Artificial Intelligence Group, Faculty of Liberal Arts and Professional Studies, York University, Toronto, Canada.
New Gener Comput. 2021;39(3-4):677-700. doi: 10.1007/s00354-021-00131-5. Epub 2021 Jun 27.
The COVID-19 pandemic resulted in a significant increase in the workload for the emergency systems and healthcare providers all around the world. The emergency systems are dealing with large number of patients in various stages of deteriorating conditions which require significant medical expertise for accurate and rapid diagnosis and treatment. This issue will become more prominent in places with lack of medical experts and state-of-the-art clinical equipment, especially in developing countries. The machine intelligence aided medical diagnosis systems can provide rapid, dependable, autonomous, and low-cost solutions for medical diagnosis in emergency conditions. In this paper, a privacy-preserving computer-aided diagnosis (CAD) framework, called Decentralized deep Emergency response Intelligence (D-EI), which provides secure machine learning based medical diagnosis on the cloud is proposed. The proposed framework provides a blockchain based decentralized machine learning solution to aid the health providers with medical diagnosis in emergency conditions. The D-EI uses blockchain smart contracts to train the CAD machine learning models using all the data on the medical cloud while preserving the privacy of patients' records. Using the proposed framework, the data of each patient helps to increase the overall accuracy of the CAD model by balancing the diagnosis datasets with minority classes and special cases. As a case study, the D-EI is demonstrated as a solution for COVID-19 diagnosis. The D-EI framework can help in pandemic management by providing rapid and accurate diagnosis in overwhelming medical workload conditions.
新冠疫情导致全球应急系统和医疗服务提供者的工作量大幅增加。应急系统正在应对大量处于病情恶化不同阶段的患者，这需要专业的医学知识来进行准确快速的诊断和治疗。在缺乏医学专家和先进临床设备的地方，这个问题会变得更加突出，尤其是在发展中国家。机器智能辅助医疗诊断系统可以为紧急情况下的医疗诊断提供快速、可靠、自主且低成本的解决方案。本文提出了一种名为去中心化深度应急响应智能（D-EI）的隐私保护计算机辅助诊断（CAD）框架，该框架在云端提供基于安全机器学习的医疗诊断。所提出的框架提供了一种基于区块链的去中心化机器学习解决方案，以协助医疗服务提供者在紧急情况下进行医疗诊断。D-EI使用区块链智能合约，利用医疗云上的所有数据来训练CAD机器学习模型，同时保护患者记录的隐私。使用所提出的框架，每个患者的数据通过平衡少数类和特殊情况的诊断数据集，有助于提高CAD模型的整体准确性。作为一个案例研究，D-EI被展示为一种新冠诊断的解决方案。D-EI框架可以通过在医疗工作量巨大的情况下提供快速准确的诊断，来帮助进行疫情管理。
