Table 3 Comparison of segmentation performance of various models on the PolyU dataset.

From: A novel approach to palm vein image segmentation combining multi-scale convolution and swin-transformer networks

Model name

Overall accuracy (OA) (%)

Dice coefficient (%)

IoU (%)

Precision (%)

Recall (%)

U-Net

95.5

90.3

84.2

92.1

89.7

ResNet

96.1

91.2

85.9

93.5

90.2

Swin-transformer

96.9

92.7

87.6

94.4

92.0

Proposed model

97.8

94.5

89.4

96.0

93.7

Back to article page