Table 5 Predictive accuracies of Double Head R-CNN model with K = 4.

From: Deep-learning-based sampling position selection on color Doppler sonography images during renal artery ultrasound scanning

Image category

Training dataset

Parameter optimization dataset

Clinical validation dataset

AO

91.5% (563/615)

78.4% (29/37)

85.1% (97/114)

NRA

94.4% (2174/2304)

92.3% (132/143)

87.5% (377/431)

RAS

87.6% (951/1086)

77.6% (52/67)

82.7% (167/202)

IRA

94.7% (2515/2656)

89.0% (146/164)

87.3% (433/496)

Back to article page