3D Point Cloud Classification On Scanobjectnn

评估指标

Mean Accuracy

OBJ-BG (OA)

OBJ-ONLY (OA)

Overall Accuracy

评测结果

各个模型在此基准测试上的表现结果

					Paper Title	Repository
OmniVec2	-	-	-	97.2	OmniVec2 - A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning	-
PointGST	-	99.48	97.76	96.18	Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learning
OmniVec	-	-	-	96.1	OmniVec: Learning robust representations with cross modal sharing	-
GPSFormer	93.8	-	-	95.4	GPSFormer: A Global Perception and Local Structure Fitting-based Transformer for Point Cloud Understanding
ReCon++	-	98.80	97.59	95.25	ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
PointGPT	-	97.2	96.6	93.4	PointGPT: Auto-regressively Generative Pre-training from Point Clouds
GPSFormer-elite	92.51	-	-	93.30	GPSFormer: A Global Perception and Local Structure Fitting-based Transformer for Point Cloud Understanding
Mamba3D	-	94.49	92.43	92.64	Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model
Mamba3D (no voting)	-	92.94	92.08	91.81	Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model
ULIP-2 + PointNeXt	91.2	-	-	91.5	ULIP-2: Towards Scalable Multimodal Pre-training for 3D Understanding
ReCon	-	95.35	93.80	91.26	Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
ULIP-2 + PointNeXt (no voting)	90.3	-	-	90.8	ULIP-2: Towards Scalable Multimodal Pre-training for 3D Understanding
ReCon (no voting)	-	95.18	93.29	90.63	Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
DeLA	89.3	-	-	90.4	Decoupled Local Aggregation for Point Cloud Learning
PCP-MAE	-	95.52	94.32	90.35	PCP-MAE: Learning to Predict Centers for Point Masked Autoencoders
PointConT	88.5	-	-	90.3	Point Cloud Classification Using Content-based Transformer via Clustering in Feature Space
Point-RAE (no voting)	-	95.53	93.63	90.28	Regress Before Construct: Regress Autoencoder for Point Cloud Self-supervised Learning
Point-FEMAE	-	95.18	93.29	90.22	Towards Compact 3D Representations via Point Feature Enhancement Masked Autoencoders
I2P-MAE (no voting)	-	94.15	91.57	90.11	Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders
ULIP + PointNeXt	88.6	-	-	89.7	ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understanding

0 of 75 row(s) selected.