Abstract
Deep learning and computer vision algorithms can deliver highly accurate and automated interpretation of medical imaging to augment and assist clinicians. However, medical imaging presents uniquely pertinent obstacles such as a lack of accessible data or a high-cost of annotation. To address this, we developed data-efficient deep learning classifiers for prediction tasks in cardiology. Using pipeline supervised models to focus relevant structures, we achieve an accuracy of 94.4% for 15-view still-image echocardiographic view classification and 91.2% accuracy for binary left ventricular hypertrophy classification. We then develop semi-supervised generative adversarial network models that can learn from both labeled and unlabeled data in a generalizable fashion. We achieve greater than 80% accuracy in view classification with only 4% of labeled data used in solely supervised techniques and achieve 92.3% accuracy for left ventricular hypertrophy classification. In exploring trade-offs between model type, resolution, data resources, and performance, we present a comprehensive analysis and improvements of efficient deep learning solutions for medical imaging assessment especially in cardiology.
Highlights
With the improved quality and accessibility of both medical imaging equipment and effective healthcare policy, medical imaging has become an increasingly critical step in modern healthcare diagnostics and procedures
We show that generative adversarial networks (GANs),[25] adapted for semi-supervised learning, can achieve better results than conventional convolutional neural networks (CNNs) in settings where labeled data is limited, achieving a test accuracy of 92.3% for Left ventricular hypertrophy (LVH)
Two different approaches for deep learning models were taken: a supervised pipeline model that performs segmentation (U-Net) before classification (CNN) and a semi-supervised generative adversarial network (GAN) for end-to-end learning. c The data, for both view and LVH classification, was split by study and no test data was utilized in training or validating the model
Summary
With the improved quality and accessibility of both medical imaging equipment and effective healthcare policy, medical imaging has become an increasingly critical step in modern healthcare diagnostics and procedures. Interpretation of medical imagery requires specialized training and is a time-intensive process. Machine learning and computer vision techniques provide an avenue to augment insights, improve accuracy, and optimize workload time for interpretation. Traditional machine learning techniques in medical imaging involve matching of features hand-engineered by domain experts, a laborious process with limited scope and effectiveness.[1,2] Recent advances in deep learning,[3,4,5] a data-driven approach, and the increasing accessibility of powerful graphical processing units (GPUs)[6,7] have made the automation of image-based diagnosis insights possible.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
