Анализ биомедицинских изображений в кардиологии на основе машинного обучения

Abstract

In this work presents a system of an automatic algorithm for analyzing images and other signals that can significantly help cardiologists in processing large amounts of data or data that require significant processing time). The analysis of heart rate variability, as well as the ECG waveform allows you to diagnose various diseases, including at an early stage, and determine specific physiological conditions, for example, the presence of physical activity now. The aim of this study was to develop a new effective method for detecting arrhythmias using image segments of ECG signals. At the first stage, image processing takes place, to identify noise in the image, data categories are allocated. At the second stage, the identification of the norm, a deviation from the norm and a critical deviation from the norm, which will allow predicting the early development of the disease and identifying diseases in patients.