Few Shot 3D Point Cloud Classification On 1

Metrics

Overall Accuracy

Standard Deviation

Results

Performance results of various models on this benchmark

Model Name	Overall Accuracy	Standard Deviation	Paper Title	Repository
PCP-MAE	97.4	2.3	PCP-MAE: Learning to Predict Centers for Point Masked Autoencoders
DGCNN	31.6	9.0	Dynamic Graph CNN for Learning on Point Clouds
I2P-MAE	97.0	1.8	Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders
ReCon++	98.0	2.3	ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
MaskPoint	95.0	3.7	Masked Discrimination for Self-Supervised Learning on Point Clouds
GPr-Net + Euc (1024)	74.4	2.0	GPr-Net: Geometric Prototypical Network for Point Cloud Few-Shot Learning
Point-MAE	96.3	2.5	Masked Autoencoders for Point Cloud Self-supervised Learning
PointGPT	98.0	1.9	PointGPT: Auto-regressively Generative Pre-training from Point Clouds
ReCon	97.3	1.9	Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
ACT	96.8	2.3	Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?
PointNet++	38.53	16.0	PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
Point-M2AE	96.8	1.8	Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training
OcCo+DGCNN	90.6	2.8	Unsupervised Point Cloud Pre-Training via Occlusion Completion
OcCo+PointNet	89.7	1.9	Unsupervised Point Cloud Pre-Training via Occlusion Completion
PointCNN	65.41	8.9	PointCNN: Convolution On $mathcal{X}$-Transformed Points
SSFSL+PointNet	63.2	10.7	Self-Supervised Few-Shot Learning on Point Clouds
GPr-Net + Hyp (512)	81.1	1.5	GPr-Net: Geometric Prototypical Network for Point Cloud Few-Shot Learning
Point-JEPA	97.4	2.2	Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud	-
GPr-Net + Hyp (1024)	80.4	0.5	GPr-Net: Geometric Prototypical Network for Point Cloud Few-Shot Learning
Point-BERT	94.6	3.1	Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling

0 of 30 row(s) selected.