Skip to main content

Table 2 Comparison of a semantic-segmentation U-Net and the path-following method for phase-unwrapping of displacement encoding with stimulated echoes (DENSE) images of the heart

From: Fully-automated global and segmental strain analysis of DENSE cardiovascular magnetic resonance using deep learning for segmentation and phase unwrapping

   MSE
Typical SNR U-Net 0.1 ± 0.2
Path-following 0.08 ± 0.08
Low SNR U-Net 0.07 ± 0.09*
Path-following 0.22 ± 0.49
  1. Mean squared error (MSE) values are reported for DENSE images with typical signal-to-noise ratio (SNR) and low-SNR
  2. *Indicates p < 0.05