Few Shot 3D Point Cloud Classification On 2

评估指标

Overall Accuracy

Standard Deviation

评测结果

各个模型在此基准测试上的表现结果

			Paper Title	Repository
ReCon++	99.5	0.8	ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
Point-JEPA	99.2	0.8	Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud
PCP-MAE	99.1	0.8	PCP-MAE: Learning to Predict Centers for Point Masked Autoencoders
PointGPT	99.0	1.0	PointGPT: Auto-regressively Generative Pre-training from Point Clouds
ReCon	98.9	1.2	Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
3D-JEPA	98.8	0.4	3D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised Representation Learning	-
point2vec	98.7	1.2	Point2Vec for Self-Supervised Representation Learning on Point Clouds
Point-RAE	98.7	1.3	Regress Before Construct: Regress Autoencoder for Point Cloud Self-supervised Learning
OTMae3D	98.7	1.2	-	-
Point-M2AE	98.3	1.4	Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training
I2P-MAE	98.3	1.3	Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders
Point-LGMask	98.1	1.4	Point-LGMask: Local and Global Contexts Embedding for Point Cloud Pre-training with Multi-Ratio Masking	-
ACT	98.0	1.4	Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?
IDPT	97.9	-	Instance-aware Dynamic Prompt Tuning for Pre-trained Point Cloud Models
Point-MAE	97.8	1.8	Masked Autoencoders for Point Cloud Self-supervised Learning
MaskPoint	97.2	1.7	Masked Discrimination for Self-Supervised Learning on Point Clouds
CrossMoCo	96.8	1.7	CrossMoCo: Multi-modal Momentum Contrastive Learning for Point Cloud	-
Point-BERT	96.3	2.7	Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling
OcCo+DGCNN	92.5	1.9	Unsupervised Point Cloud Pre-Training via Occlusion Completion
OcCo+PointNet	92.4	1.6	Unsupervised Point Cloud Pre-Training via Occlusion Completion

0 of 30 row(s) selected.