Skip to main content

Table 3 Spatial overlap results demonstrating the algorithm’s precision. The algorithm was run five times on two arbitrarily selected data sets

From: Semi-automated left ventricular segmentation based on a guide point model approach for 3D cine DENSE cardiovascular magnetic resonance

  Dice Jaccard False positive False negative
Iteration Set 1 Set 2 Set 1 Set 2 Set 1 Set 2 Set 1 Set 2
1 0.9225 0.9083 0.8562 0.8321 0.0162 0.0856 0.1299 0.0586
2 0.9064 0.9187 0.8289 0.8496 0.0608 0.0776 0.1331 0.0628
3 0.9068 0.9160 0.8295 0.8450 0.0481 0.0972 0.1306 0.0729
4 0.9183 0.9261 0.8490 0.8623 0.0719 0.0833 0.0899 0.0658
5 0.9115 0.9370 0.8374 0.8814 0.0656 0.0202 0.1158 0.1008
  1. An arbitrary slice at an early systolic cardiac phase is selected and was manually contoured by a single Operator. The spatial overlap measures were calculated after each experiment. The results for each data set show the consistency of the algorithm.