Table 3.

Summary of reviewed works on supervised and unsupervised machine learning.

Subcategory	Related works	Year	Technique	Filter	Database	Evaluation metric
USML	[148]	2012	Clustering	2-D median	MIAS	90.0% sensitivity and 78.0% specificity
USML	[140]	2012	Microcalcification clusters		BNHMJ	91.4% segmentation accuracy, false positive 96.5%
USML	[142]	2013	FCM clustering	Morphological	MIAS
USML	[138]	2013	Microcalcification clusters		DDSM	93.2% positive rate and 0.73 false positive
USML	[147]	2014	k-means	5 × 5 median	MIAS	94.4% sensitivity
USML	[145]	2015	Fuzzy c-means		MIAS	83.3% for class 1, 75.0% class 2, and 80.0% class 3 accuracy
USML	[149]	2017	FCM clustering		MIAS	86.2% sensitivity, 96.4% specificity, and 94.6% accuracy
USML	[139]	2018	MC clusters	Morphological	DDSM and MIAS	94.48% classification accuracy for DDSM and 100.0% for MIAS
USML	[136]	2018	Fuzzy c-means clustering		MIAS	98.82% detection
USML	[137]	2018	c-means clustering		MIAS	98.1% accuracy
USML	[141]	2018	Classic and fuzzy morphology	Gaussian	MIAS	0.86 Dice, 66.0% recall and 20% precision
USML	[144]	2018	c-means	LoG	MIAS and PHP	95.0% accuracy PHP and 94.0% MIAS
USML	[143]	2018		Morphological	DDSM and MIAS	98.0% accuracy for MIAS and 97.0% for DDSM accuracy
USML	[135]	2018	Hierarchical k-means clustering		DDSM	38.8% accuracy and 61.1% testing error
USML	[146]	2018	MC clusters	Morphological	DDSM	96.57% sensitivity and 94.25% accuracy
SML	[155]	2011	MLP		DDSM	68.2% sensitivity and 8.7% false positive per image
SML	[156]	2012	ELM		MIAS	81.10% of accuracy
SML	[150]	2015	Structure SVM		DDSM and INbreast	87.0% Dice
SML	[152]	2015	SSVM and CRF		DDSM and INbreast	93.0% accuracy using CRF and 95.0% accuracy using SVM
SML	[153]	2015	SVM	Median filter	SSPS	96.0% correlation
SML	[151]	2016	GGD and Bayesian back propagation		MIAS	97.08% detection for GGD and 97.0% for Bayesian
SML	[154]	2017	CRF and SSVM		DDSM and INbreast	10.0% loss