Evaluation of a Nonrigid Motion Compensation Technique Based on Spatiotemporal Features for Small Lesion Detection in Breast MRI

Table 5

Areas under the ROC curves for scaling index (SI) method using FLDA. The rows represent the motion compensation as given by Table 1. Numbers in boldface show the best results.

Feature type	Area under the ROC curve (%)
	01	02	03	04	05	06	07	08	09
SI max.	74.6	66.6	64.3	69.3	62.8	80.3	69.1	60.9	62.8
SI mean	80.3	79.8	79.6	80.0	78.6	80.5	77.1	79.6	77.5
SI st. dev.	52.7	81.5	73.3	70.8	74.4	61.1	67.4	79.2	72.5
SI entropy	70.8	71.4	75.0	72.7	71.6	68.9	65.1	75.0	76.5