Abdominal Tumor Characterization and Recognition Using Superior-Order Cooccurrence Matrices, Based on Ultrasound Images

Table 3

The values of the accuracy parameters obtained by using the final set of relevant textural features appropriate for the differentiation between HCC and the benign liver tumors.

Classifier	Recognition rate	TP rate	TN rate	AUC	Time
SVM	83.16%	78.9%	87.4%	83.2%	81.32 s
MLP	82.66%	78.4%	86.9%	89.9%	62.27 s
AdaBoost + SVM	83.21%	79%	88.3%	83.9%	89.22 s
AdaBoost + MLP	83.66%	80.4%	86.9%	84.3%	89.16 s