Research Article

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

Table 6

Areas under the ROC curves for spatio-temporal features using FLDA. The rows represent the motion compensation as given in Table 1.

Feature typeArea under the ROC curve (%)
010203040506070809

RSIE st. dev. ( 3 ) 56.752.555.052.348.152.555.558.455.3
RSIE st. dev. ( 4 ) 64.965.868.562.863.766.666.869.564.5
RSIE st. dev. ( 5 ) 64.770.670.064.968.369.765.866.869.5

RSIE entropy ( 3 ) 80.9 84.285.3 83.079.484.581.381.583.4
RSIE entropy ( 4 ) 77.987.481.980.776.783.878.877.378.4
RSIE entropy ( 5 ) 74.6 79.881.781.773.381.5 76.376.973.5

Contour RSIE mean ( 3 ) 54.451.752.752.354.055.752.752.755.5
Contour RSIE mean ( 4 ) 63.756.962.859.261.159.559.559.059.9
Contour RSIE mean ( 5 ) 62.264.960.368.568.762.663.962.666.6

Contour RSIE st. dev. ( 3 ) 55.357.458.052.155.758.055.057.457.8
Contour RSIE st. dev. ( 4 ) 58.459.957.156.761.360.758.656.962.4
Contour RSIE st. dev. ( 5 ) 63.763.267.660.565.158.658.266.264.3

Contour RSIE entropy ( 3 ) 77.581.377.174.276.977.174.875.474.4
Contour RSIE entropy ( 4 ) 83.284.582.879.877.984.977.581.379.6
Contour RSIE entropy ( 5 ) 80.581.378.280.772.979.678.277.578.4