Novel and efficient classification of cardiovoscular abnormalities using machine learning and signal processing techniques

Abstract

With the invention of medical expert systems, the demand for efficient innovative techniques in signal processing to detect abnormalities is ever increasing for identifying cardiovascular diseases. Recently, diseases related to the heart are increasing exponentially with time. An electrocardiogram signal gives information about the heart electrical activity of the person. Heart condition can be easily monitored, with the help of Electrocardiogram (ECG) signals obtained from a person. newlineThe first research work objective is to perform ECG signal processing and to apply machine learning tools in developing best classifier to classify ECG data into normal or abnormal cases. Further to design and implement hardware module named as VH-Doctor which can help people to treat heart diseases. K nearest neighbor (KNN) classifier based cardiovascular disease abnormal detection is proposed. In this research, ECG signal processing, Feature extraction and KNN classifier are performed and achieve the highest accuracy of 99% better than other machine learning algorithms. The major objective of this research is to offer medical services to people in remote villages at low cost. People in villages and remote areas do not have facilities to get treated by a medical expert. This research provides them an opportunity to get medical advice through the virtual environment called the VH-Doctor machine. It is a virtual environment Heart doctor and reduces human effort in testing and treating heart diseases at the initial stages. The patients are treated and diagnosed with the help of machines without human effort. Biomedical sensors, ARM processor and FPGA used to detect, test, analyze and display normal or abnormal cases. newline