Evaluation of transfer learning in deep convolutional neural network models for cardiac short axis slice classification

Abstract

In computer-aided analysis of cardiac MRI data, segmentations of the left ventricle (LV) and myocardium are performed to quantify LV ejection fraction and LV mass, and they are performed after the identification of a short axis slice coverage, where automatic classification of the slice range of interest is preferable. Standard cardiac image post-processing guidelines indicate the importance of the correct identification of a short axis slice range for accurate quantification. We investigated the feasibility of applying transfer learning of deep convolutional neural networks (CNNs) as a means to automatically classify the short axis slice range, as transfer learning is well suited to medical image data where labeled data is scarce and expensive to obtain. The short axis slice images were classified into out-of-apical, apical-to-basal, and out-of-basal, on the basis of short axis slice location in the LV. We developed a custom user interface to conveniently label image slices into one of the three categories for the generation of training data and evaluated the performance of transfer learning in nine popular deep CNNs. Evaluation with unseen test data indicated that among the CNNs the fine-tuned VGG16 produced the highest values in all evaluation categories considered and appeared to be the most appropriate choice for the cardiac slice range classification.

Classification of Cardiomyopathies from MR Cine Images Using Convolutional Neural Network with Transfer Learning

“…This model seems to be well generalizable, for example, with images obtained after gadolinium injection (mostly short axis), comprising pleural or pericardial effusion, or coming from different kinds of MR scanners. Other base models tested did not provide better results, and this is in line with results reported for cardiac short-axis slice range classification [ 20 ]. In a large, multicenter study, Betancur et al [ 9 ] used standard Convnet with three feature extraction units for prediction of obstructive coronary artery disease by SPECT.…”

Transferable, Deep-Learning-Driven Fast Prediction and Design of Thermal Transport in Mechanically Stretched Graphene Flakes

“…The P matrix and fingerprint of the piled graphene structures in the custom data set will be treated as inputs to the training of the DNN, similar to that in Figure a. Different from a completely training from scratch of the DNN in Figure a, we introduce fine-tuned layers that allow transferring the “knowledge” of the trained DNN from the primitive data set − and updating the trained DNN with the geometric features of piled graphene structures from the custom data set. With the updated DNN, an expanded databank that includes information from both primitive and custom data sets will be established.…”

Classification of Gliomas and Germinomas of the Basal Ganglia by Transfer Learning

“…Currently, there is no well-acknowledged method for adding pseudo color to medical images. Previously reported method includes linear color conversion from grayscale to a color map (29), triplicate the grayscale channel to synthesize color image (30,31), concatenating three independent slices from one or cross different series (planes) (32)(33)(34)(35). In this study, we thoroughly benchmarked these methods.…”