EchoPWC-Net is a novel pipeline for myocardial function imaging in echocardiography using deep learning. This modified PWCNet motion estimation network can perform on the same level, or even better, compared to other known methods when training on simulated ultrasound data. This pipeline is able to estimate longitudinal strain automatically in a prospective nature.
Creator
Information
Pediatrics or Adult
Adult
Speciality
Cardiology
Modality
Echocardiogram
Training
“FlyingChairs2D (total of 22872 images of 3D chair models) and FlyingThings3D (25000 stereo images sampled from a 3D scene)”