Table 1 Comparison of the linear SVM classification on ModelNet40

From: PointMoment: a mixed-moment self-supervised learning approach for 3D Terracotta Warriors

Method

ModelNet40

3D-GAN42

83.3

Latent-GAN17

85.7

SO-Net43

87.3

FoldingNet18

88.4

MRTNet44

86.4

3D-PointCapsNet45

88.9

MAP-VAE46

88.4

DepthContrast37

85.4

Jigsaw47+PointNet

87.3

Rotation11+PointNet

88.6

OcCo19+PointNet

88.7

STRL24+PointNet

88.3

PointMoment-all(Ours)+PointNet

88.8

PointMoment-singler(Ours)+PointNet

88.8

Jigsaw47 + DGCNN

90.6

Rotation11 + DGCNN

90.8

STRL24 + DGCNN

90.9

OcCo19 + DGCNN

89.2

PointMoment-all(Ours)+DGCNN

90.9

PointMoment-singler(Ours)+DGCNN

91.0

  1. The linear classifier is fitted on the training set of ModelNet40 using the pre-trained model, and the model performance is evaluated on the test set.
Back to article page