Skip to main content


Table 1 Specificity and sensitivity are shown in brackets (in that order) for the cross-validation experiments.

From: Optimal patient classification via statistical decomposition of a 3D left-ventricular atlas

PCA (53.8,54.4) (77.5,75.6) (89.4,87.9)
IPCA (96.6,91.2) (96.2,98.0) (100,99.0)
EF    (82.7,93.2)
EDV (81.6,92.4)   
ESV   (79.2,98.3)  
LVM (67.1,70.1)   
  1. Columns indicate the type of information used by the classifier; rows indicate which classifier is used. Top two rows are different statistical decomposition techniques of shape parameters (preserving only 1 dimension) and bottom four rows are traditional clinical heart-failure indicators. In the case of ED&#8722ES, PCA and IPCA include information on regional wall motion.