A High-Precision Classification Method of Mammary Cancer Based on Improved DenseNet Driven by an Attention Mechanism

Table 7

Comparison of evaluation results of multiclass classification methods in different classification tasks.

Methods	Classification task	Average accuracy (%)	Mean accuracy (%)	Average recall rate (%)	Average value (%)
DenseNet121	Binary classification	91.05	91.34	92.01	91.76
	Multiclass classification	84.01	83.73	85.98	84.84
MFSCNet A	Binary classification	99.48	99.61	99.23	99.42
	Multiclass classification	96.71	97.02	96.54	96.78
MFSCNet B	Binary classification	98.73	97.34	98.84	97.65
	Multiclass classification	93.14	93.42	92.95	93.18
MFSCNet C	Binary classification	99.06	98.86	98.48	98.67
	Multiclass classification	92.23	91.34	93.18	92.25