Fus2Net: a novel Convolutional Neural Network for classification of benign and malignant breast tumor in ultrasound images

Table 2 The performance metrics after using image augmentation, L2 regularization, and dropout

	Accuracy	Sensitivity	Specificity	Precision	AUC
Image augmentation	88.81%	92.05%	82.58%	85.37%	0.91
Image augmentation and L2 regularization and dropout	93.25%	94.19%	88.57%	91.94%	0.97

The bold value in the table indicates this method (in row) outforms others regarding this specific metric (in column)

ISSN: 1475-925X