Hybrid Inception v3 XGBoost Model for Acute Lymphoblastic Leukemia Classification

Table 2

Evaluation metrics of different pretrained CNN models from stage I of training. These metrics are reported on the test set after fine-tuning on the train set.

Model	F1 score	Recall	Precision	Accuracy	AUC
AlexNet	0.889	0.894	0.901	0.894	0.832
DenseNet 121	0.871	0.869	0.876	0.869	0.861
ResNet 18	0.917	0.917	0.919	0.917	0.908
VGG 16	0.921	0.924	0.927	0.924	0.880
SqueezeNet	0.930	0.932	0.936	0.932	0.891
MobileNet v2	0.958	0.958	0.958	0.958	0.953
Inception v3	0.979	0.979	0.979	0.979	0.981