A portable medical device for detecting diseases using Probabilistic Neural Network. (January 2022)

• Record Type: Journal Article
• Authors: Moradkhani, Ali; Broumandnia, Ali; Mirabedini, Seyed Javad
• Is Part Of: Biomedical signal processing and control. Volume 71(2022)Part A
• Publication Date: 2022-01
• Subjects: Probabilistic Neural Network (PNN) -- Cancer detection -- Bioelectrical impedance analysis (BIA) -- Stress test -- Electrocardiogram (ECG); Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2021.103142
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Many people die from lung and breast cancer or heart diseases every year due to a lack of taking diagnosis tests. These diseases are usually diagnosed when they have reached an advanced stage that is why it is so difficult or impossible to treat them. The early diagnosis of the disease requires periodically regular tests. However, it is almost unattainable for patients to go to a medical center to have tests frequently, because of the rush of everyday life. Two groups of nine women and nine men who had diseases such as breast or lung cancer, heart disease, or those who had the potential to suffer from these diseases, were selected for this research. A wearable hardware is designed and embedded in a belt to measure, collect, and transfer vital data such as ECG, body impedance, and some other body's information to a server for cloud processing, using the patient's cellphone as a communication interface. On the server, the data collected from the patients is compared to the stored data of other patients who have a similar disease. This comparison is made by a devised algorithm, which deploys the Probabilistic Neural Network as an inference, machine learning system to determine the course of disease changes or the process of disease recovery. The results show that the hardware and the algorithm applied in this study perform better at diagnosing lung cancer than breast cancer and heart disease.