Skip to main content

Table 3 Input types experiments setting and result

From: Deep learning based approaches from semantic point clouds to semantic BIM models for heritage digital twin

Input type

mIoU (%)

Build.

Tra.

Cars

Box (4 m × 4 m)

37.78

57.44

73.98

3.16

Box (10 m × 10 m)

42.00

65.28

81.27

1.15

Sphere (2 m-radius)

43.16

66.52

77.08

1.81

Sphere (4 m-radius)

46.93

70.81

87.77

6.41