Meeting Banner
Abstract #0634

Automated Left Ventricular Volumetric Quantitation from Short-axis CMR Images with Machine Learning using a Deep Convolutional Neural Network

James W Goldfarb1, Jie J Cao1, and Julian de Wit2

1St Francis Hospital, Roslyn, NY, United States, 2DWS Systems, Hoek van Holland, Netherlands

Automatic segmentation of the LV bloodpool using deep learning with a convolutional neural network is a promising, accurate and efficient method for segmentation of cardiac MR images. Although there were a few cases with inaccurate results, "big fails", accuracy is high, R2=0.93 and ejection fraction error ~4%. In the future it may provide a customizable, fast and accurate method for comprehensive evaluation of cardiac MR images.

This abstract and the presentation materials are available to members only; a login is required.

Join Here