Skip to main content

Table 3 Difference in net flow between manual segmentation in relation to machine learning and conventional (commercially available) automated segmentation

From: Machine learning derived segmentation of phase velocity encoded cardiovascular magnetic resonance for fully automated aortic flow quantification

 

Net Flow

Absolute Difference (|manual – method|)

Correlation

Manual

81.5 ± 24.2 mL

  

Machine Learning

80.5 ± 23.7 mLa

1.85 ± 1.80 mLb

y = 1.01x + 0.16

r2 = 0.99, p < 0.001

Conventional

80.1 ± 23.2 mLa

3.33 ± 3.18 mLb

y = 1.02x – 0.31

r2 = 0.96, p < 0.001

  1. aBoth p < 0.01 (segmentation method vs. manual)
  2. bp < 0.01 (machine learning vs. conventional segmentation in terms of MAD)
Back to article page

Journal of Cardiovascular Magnetic Resonance

ISSN: 1532-429X

Contact us