Modelling of biosignal based decision making model for intracranial haemorrhage diagnosis in IoT environment

• Published in: Expert systems Vol. 39; no. 7
• Main Authors: Hilal, Anwer Mustafa, Alabdan, Rana, Othman, Mohamed Tahar Ben, Hassine, Siwar Ben Haj, Al‐Wesabi, Fahd N., Rizwanullah, Mohammed, Yaseen, Ishfaq, Motwakel, Abdelwahed
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01-08-2022
• Subjects: Algorithms; biosignals; Classification; CMOS; Computed tomography; Data acquisition; Datasets; Decision making; deep learning; Diagnosis; Feature extraction; Health care; Hemorrhage; Image segmentation; Internet of Things; intracranial haemorrhage; Machine learning; Medical imaging; Neural networks
• ISSN: 0266-4720; 1468-0394
• DOI: 10.1111/exsy.12964

In recent times, the Internet of Things (IoT) in healthcare is a novel and promising emerging area that offers a number of benefits to users and healthcare professionals, enabling real time monitoring of diseases, convenience, and ease of use. With respect to aid the patients by the use of IoT with biosignals enabled solutions, recent researches make use of machine learning (ML) techniques for decision making, particularly haemorrhage diagnosis. In this view, this paper presents an intelligent intracranial haemorrhage (ICH) diagnosis using biosignals (IICHD‐BS) in IoT environment. The IICHD‐BS technique performs data acquisition process using two sensors namely complementary metal oxide semiconductor sensor and ESP8266 Wi‐Fi module. They are employed to collect CT images of the CT scanner and transforms them into electrical signals for storing them in the server. Besides, the IICHD‐BS technique employs optimal region growing based segmentation approach for detecting the infected brain regions in the CT images. In addition, EfficientNet based feature extraction and functional link neural network (FLNN) based classification approach is used for detecting and classifying the existence of ICH. For experimental validation, a set of two benchmark dataset ICH datasets are used and the experimental outcomes are evaluated with respect to different measures. The simulation outcomes demonstrated the improved performances of the proposed algorithm over the current state of art techniques.