China
Africa
Explore chapters and articles related to this topic
Deep Learning for Medical Dataset Classification Based on Convolutional Neural Networks
Published in R. Sujatha, S. L. Aarthy, R. Vettriselvan, Integrating Deep Learning Algorithms to Overcome Challenges in Big Data Analytics, 2021
In medical image studies, spatial and temporal data are widely used with the help of AI and Deep Learning. There are many models of cardiac imaging, such as MRI, CT, Ultrasound, and others. Using these modalities, the classification and detection of heart diseases are identified with the visualization effects of the structures. While handling the classification and segmentation of the cardiovascular anatomy, cardiovascular imaging is mainly focused on AI and Deep Learning-based strategies. Also, the imaging modalities are based on the morphologies in the Left/Right Ventricle (LV/RV), the coronary tree, and the aorta, the valve plane. In Cardiac Magnetic Resonance Imaging (CMRI), Deep and Machine Learning algorithms can be used for the identification and detection of cardiac functions and for the accuracy of ventricular volumes. In echocardiography (ECHO), the standard views are used for the detection and the segmentation of the LV (Bello et al., 2019).
Analysis of the four heart sounds statistical study and spectro-temporal characteristics
Published in Journal of Medical Engineering & Technology, 2020
This study gives a classification of heart sound components and murmurs, and their relation to cardiac structures and hemodynamic variables. Whereas auscultation still holds its position as a diagnostic tool for general physicians and cardiologists, phonocardiography in its classical form has lost interest mostly as a result of the availability of techniques as Doppler echocardiography and cardiac imaging methods, which give more direct information on cardiac structures and hemodynamic variables. The historical value of conventional phonocardiography has to be stressed. On the other hand, recording and processing of heart sounds continue to be subjects of scientific research and remain beneficial for training and for supporting of diagnosis of physicians. Electronic stethoscopes coupled to a laptop and connected to the internet for automated or for remote diagnosis by a specialist may grow in importance in the next few years.
Artificial intelligence: improving the efficiency of cardiovascular imaging
Published in Expert Review of Medical Devices, 2020
Andrew Lin, Márton Kolossváry, Ivana Išgum, Pál Maurovich-Horvat, Piotr J Slomka, Damini Dey
Echocardiography remains the most widely used cardiac imaging modality. The increasing uptake of hand-held ultrasound devices and focused scanning protocols have enabled the rapid, point-of-care evaluation of cardiac structure, function and hemodynamics [13]. However, performance of echocardiographic examinations is operator-dependent and intensive training is required for proficiency. The automation of image acquisition may eliminate some of these training and skill maintenance requirements. Commercial vendors have developed DL-based real-time guidance software to guide untrained providers in acquiring standard echocardiographic views, whereby the operator optimizes an accuracy signal that indicates how close they are to the desired view [14]. These DL models can automatically analyze the acquired image dataset to identify and label heart anatomy, then slice standard views for presentation. This process reduces inter-operator variability, as the optimal views are selected based on thousands of deep learned examinations representing the spectrum of anatomical variations. Recently, a CNN used to identify 8 standard 3D echocardiographic views achieved an accuracy of 92% compared to ground truth training data labeled by clinicians [15]. Such automation of image acquisition may enable non-cardiologists to perform echocardiography as a first-line diagnostic tool in acute settings, and also facilitate the training of cardiac sonographers.
Using particle systems for mitral valve segmentation from 3D transoesophageal echocardiography (3D TOE) - a proof of concept
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Patricia Lopes, Paul Van Herck, Eefje Verhoelst, Roel Wirix-Speetjens, Jan Sijbers, Johan Bosmans, Jos Vander Sloten
The results obtained through the described method were tested by comparing its output with a manual process. The manual segmentation was initiated by a biomedical engineer with experience in the Mimics software and 3 years of echocardiographic image analyses. For each of the selected diastolic and systolic phases, the aortic valve centre, the mitral valve centre and a point in the left ventricle were indicated to generate the long-axis view, which was subsequently rotated at 20° intervals. The mitral annulus was selected on each of these rotational planes, and for the mid-diastolic images, the operator additionally indicated the tip of the leaflets’ free edge to define the valve opening. Subsequently, the operator segmented the leaflets by manually placing points on the images as considered required. A thin plate spline method was used to concurrently fit a surface to the placed points, allowing the operator to visually assess the progress of the segmentation. Once the leaflet delineation was concluded, the regions of the thin plate spline surface not circumscribed by the annulus and, in the case of the diastolic phases, by the curve describing the free leaflet edge, were excluded (Figure 1). The resulting surfaces were subsequently reviewed by a cardiologist with over 10 years of cardiac imaging experience, including echocardiography and cardiac computed tomography (CT). Corrections to the initial segmentation were performed using an interactive contour editing tool, also available in Mimics. The output of the manual approach was compared with the result of the proposed method by measuring the root mean square (RMS) difference between the two meshes, taking successively each surface as reference, and calculating the mean of the two RMS values.