Multiple Active Contours Guided by Differential Evolution for Medical Image Segmentation

Table 1

Average similarity measure with Hausdorff distance, Jaccard index, and Dice index among the regions segmented by classical active contour model (ACM), interactive Tseng method, our proposed method (MACDE), and the regions outlined by two experts from the set of CT images.

Comparative studies	Distance/similarity measure
	Hausdorff (𝐻)	Jaccard index (𝐽)	Dice index (𝐷)
ACM versus Expert 1	4.0	0.3548	0.5238
ACM versus Expert 2	3.0	0.5272	0.6904
Tseng versus Expert 1	2.236	0.8260	0.9047
Tseng versus Expert 2	2.8284	0.7872	0.8809
MACDE versus Expert 1	3.0	0.8666	0.9285
MACDE versus Expert 2	1.4142	0.9090	0.9523