Multi-Features-Based Automated Breast Tumor Diagnosis Using Ultrasound Image and Support Vector Machine

Table 5

The comparison of classification methods (CFs—characteristic features, DLFs—deep learning features, and MFs—multi-features).

Models	Accuracy	Precision	Sensitivity	Specificity	F1-score
MF + SVM	0.925	0.905	0.95	0.905	0.927
CF + SVM	0.875	0.895	0.85	0.895	0.827
DLF + SVM	0.8	0.875	0.7	0.875	0.778
[3]	0.901	NA	0.935	0.832	NA
[7]	0.83	NA	NA	0.824	NA
Inception V3 [36]	0.78	NA	0.77	0.78	NA
VGG19 [36]	0.82	0.70	0.70	0.78	NA
[37]	0.8667	NA	0.9245	0.7838	NA

The bold values indicate that the result of the proposed method is better than that of other classification methods.