Deep Transfer Learning for Automatic Prediction of Hemorrhagic Stroke on CT Images

Table 2

Overview of related work on classification of ICH using CT scans.

Author	Dataset	Method	Performance (%)
Wei et al. [30]	212	ResNet	Accuracy 95.3
Arbabshirani et al. [41]	46583	CNN	Accuracy: 84
Chang et al. [43]	536255	3D/2D mask R-CNN	Sensitivity: 95
Dhar et al. [44]	224	Deep learning	Accuracy: 98
McLouth et al. [45]	1192	Deep learning	Accuracy: 95.6
Danfeng et al. [48]	1176	Fully CNN	Accuracy: 95
Patel et al. [49]	1554	CNN-RNN	Accuracy: 96
Li et al. [50]	159	U-Net	Accuracy: 98.5
Proposed model	1164	ResNet-50 and dense layer	Accuracy: 99.6; Sensitivity: 99.4