Table 3 The efficiency analysis results of each method on the SemanticKITTI test set. Please note that all results are from the corresponding literature.

From: Multi-scale sparse convolution and point convolution adaptive fusion point cloud semantic segmentation method

Method

Params (M)

MACs (G)

Latency (ms)

mIoU (%)

SqueezeSeg-V3

26.2

515.2

238

55.9

PolarNet

13.6

135.0

62

54.3

SPVNAS

12.5

73.8

259

61.0

SalsaNext

14.7

109.3

103

59.5

LG-Net

15.6

72.4

115

56.3

PCSCNet

15.4

80.9

245

62.7

Our Method

17.3

92.7

287

65.6