AI-Cardio project: Artificial intelligence system for diagnosing and predicting myocardial infarction based on echocardiography provides the ability to automatically calculate a number of important indicators of heart health such as ejection fraction, total longitudinal strain of the left ventricle. On that basis, the software supports doctors in diagnosing and predicting myocardial infarction based on echocardiography.
Algorithms that automatically analyze echocardiographic images using the latest Artificial Intelligence methods, specifically image recognition methods based on Deep Learning techniques, were developed by the research team to identify animated gif of the heartbeat. In order for the algorithms to work accurately, doctors painstakingly labeled to create data sets for thousands of echocardiograms, on different types of echocardiography machines. Based on the results of identification and analysis of ultrasound images, the software automatically estimates important indicators of the heart allowing diagnosis and prognosis of myocardial infarction. With this approach, the project automates some functions that still have to be performed manually or semi-automatically on echocardiography machines, which requires doctors with good technical expertise and is currently quite expensive. This automatic feature helps lower-level hospitals screen patients for myocardial infarction better and more effectively.
