Table 12.

A summary of CNN applications in medical image segmentation surveyed in Sect. 5

References	Organ	Dataset	Comparison against	Performance metric
[8]	MS lesion	ISBI 2015 and MICCAI 2008 datasets	Five publicly available methods	DSC, TPR and FPR. Achieved better TPR and FPR, but less DSC values
[36]	Breast	Dutch breast cancer screening dataset	State of the art methods	AUC
[72]	Breast	The Cancer Genome Atlas breast cancer dataset	Three texture classification methods namely RPLSVM, LBPLSVM and THLB	Accuracy, efficiency and scalability and F-score of $85\%$
[77]	Nucleus	Brain tumor, pancreatic NET and breast cancer	SVM, RF and DBN	Precision, recall and F-score
[83]	Heart	CCTA scans of 60 patients	Manual segmentation	Sensitivity $95\%$ and specificity $96.6\%$ DSC $85\%$ and mean absolute surface distance $85\%$
[81]	Eye glaucoma	DRISHTI-GS, Messidor	Manual	Precision, recall, F-score