Cardiovascular Image Segmentation in Computed Tomography Angiography ImagesUsing Deep Learning Approaches

Mohd Osama; Rajesh Kumar

Abstract

In present time, the cardiovascular disease is one of the common causes of mortality in human. In field of medical science, Heart angiography is one of the processes to testing of heart disease. Heart angiography identifies the abnormality in heart vessels. There are mainly two approaches to identify the heart disease. Former approach is the invasive and latter one is the non-invasive approaches. Invasive process is a painful diagnostic procedure that provides important information of heart structure and function of the heart. In Invasive procedure, a catheter has been penetrated to get X-rays of the heart's arteries (i.e. coronary arteries) called coronary Angiography or Arteriography. In noninvasive painless procedure cardiovascular diagnostic testing includes the complete spectrum of heart. Noninvasive process has been done by highly skilled doctors and nurses, that having the knowledge to latest technology. In this paper a segmentation approach of multiple structure of CT angiographic images using U-net approach of deep learning method has been proposed. U-net architecture has been applied in CT angiography and experimental results had evaluated in term of Precision, Recall, F1 Score and accuracy.

Abstract

Keywords