Table 3 Scene segmentation result on S3DIS.

From: Frame points attention convolution for deep learning on point cloud

Methods

OA

instance mIoU

PointNet19

73.9

41.1

PointNet++20

83.7

53.1

SPG52

86.4

58.0

PointCNN30

85.9

58.3

HAPGN28

85.8

–

DPC53

86.8

61.3

DPAM54

86.1

60.0

ADConvnet-SAGC55

87.5

60.1

GACNet27

87.8

62.9

SPT37

89.5

68.9

LCPFormer36

90.8

70.2

PT34

90.8

70.4

FPAC

91.0

70.5

  1. The best performance for each metric is shown in bold..
Back to article page