Skip to main content

Table 3 Dice’s coefficient (DC) values of different burn sizes in different models

From: Burn image segmentation based on Mask Regions with Convolutional Neural Network deep learning framework: more accurate and more convenient

Burn sizes

Model name

R101FA (our method)

IV2RA

R101A

%TBSA < 5%

83.6

86.48

86.63*

5% < %TBSA< 20%

86.13

88.06*

84.85

%TBSA> 20%

81.27*

72.33

74.13

  1. *The highest average DC value of this burn size in different models
  2. R101FA residual network-101 with atrous convolution in feature pyramid network, IV2RA inceptionV2-residual network with atrous convolution, R101A residual network-101 with atrous convolution