Key result
Deep neural networks for automatic 2D LV segmentation demand tedious and time-consuming expert data annotation.
Deep learning shows promise for automatic 2D left ventricle segmentation in echocardiography, though it requires large annotated datasets.
High annotation burden limits clinical adoption of automated LV segmentation; leaves open data-efficient DL approaches in echocardiography.
Automatic segmentation of the left ventricle (LV) can become a useful tool in echocardiography, for instance to provide automatic ejection fraction measurements or to initialize deformation imaging algorithms. Deep neural networks have recently shown very promising results for improving image classification and segmentation. These methods learn using only a set of input and output data, but require a large and representative amount of annotated data to be successful. This means an expert has to draw the LV border in potentially thousands of images, which is highly tedious and time consuming.
No takes yet. Share an insight, caveat, or question.
Smistad et al. (2017) studied Left ventricle segmentation. Deep neural networks was evaluated. Deep neural networks for automatic 2D left ventricle segmentation require a large amount of annotated data, which is highly tedious and time-consuming for experts to produce.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: