HyperAIHyperAI

Command Palette

Search for a command to run...

Console

Few Shot 3D Point Cloud Classification On 1

Metrics

Overall Accuracy
Standard Deviation

Results

Performance results of various models on this benchmark

Paper Title
ReCon++98.02.3ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
PointGPT98.01.9PointGPT: Auto-regressively Generative Pre-training from Point Clouds
3D-JEPA97.62.03D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised Representation Learning
PCP-MAE97.42.3PCP-MAE: Learning to Predict Centers for Point Masked Autoencoders
Point-JEPA97.42.2Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud
Point-LGMask97.42.0Point-LGMask: Local and Global Contexts Embedding for Point Cloud Pre-training with Multi-Ratio Masking
ReCon97.31.9Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
Point-RAE97.31.6Regress Before Construct: Regress Autoencoder for Point Cloud Self-supervised Learning
IDPT97.3-Instance-aware Dynamic Prompt Tuning for Pre-trained Point Cloud Models
OTMae3D97.22.3-
I2P-MAE97.01.8Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders
point2vec97.02.8Point2Vec for Self-Supervised Representation Learning on Point Clouds
ACT96.82.3Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?
Point-M2AE96.81.8Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training
Point-MAE96.32.5Masked Autoencoders for Point Cloud Self-supervised Learning
MaskPoint95.03.7Masked Discrimination for Self-Supervised Learning on Point Clouds
Point-BERT94.63.1Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling
CrossMoCo93.84.5CrossMoCo: Multi-modal Momentum Contrastive Learning for Point Cloud
OcCo+DGCNN90.62.8Unsupervised Point Cloud Pre-Training via Occlusion Completion
OcCo+PointNet89.71.9Unsupervised Point Cloud Pre-Training via Occlusion Completion
0 of 30 row(s) selected.