A Lightweight Hybrid Dilated Ghost Model-Based Approach for the Prognosis of Breast Cancer

Table 4

The dice index, intersection over union, and pixel accuracy comparisons of the proposed model and traditional methods.

S.N.	Methods	Dice index	Intersection over union	Pixel accuracy
1	ResNet	0.5989	0.4893	0.6813
2	Grab cut	0.4663	0.3491	0.6220
3	Level set	0.5980	0.4587	0.6758
4	DenseNet	0.5464	0.4322	0.6440
5	The proposed model	0.7697	0.6041	0.7983