Skip to main content
. 2022 Nov 22;10:1056226. doi: 10.3389/fpubh.2022.1056226

Table 3.

Comparative experiments of Dropblock and Dropout on four datasets.

Datasets Models AUC F1 IOU MCC
DRIVE Baseline+Dropout 0.9841 0.8147 0.6873 0.7994
Baseline+Dropblock 0.9860 0.8171 0.6908 0.8012
U-Net 0.9849 0.8170 0.6907 0.8007
CHASE DB1 Baseline+Dropout 0.9848 0.7976 0.6634 0.7844
Baseline+Dropblock 0.9892 0.8101 0.6809 0.7973
U-Net 0.9873 0.7989 0.6652 0.7853
STARE Baseline+Dropout 0.9802 0.8076 0.6781 0.7945
Baseline+Dropblock 0.9885 0.8238 0.7010 0.8115
U-Net 0.9813 0.8026 0.6709 0.7887
IOSTAR Baseline+Dropout 0.9874 0.8019 0.6694 0.7886
Baseline+Dropblock 0.9884 0.8152 0.6879 0.8009
U-Net 0.9873 0.8104 0.6813 0.7967

Bold values is the highest scores for the metrics.