SOTAVerified

Training-free 3D Point Cloud Classification

Evaluation on target datasets for 3D Point Cloud Classification without any training

Papers

Showing 18 of 8 papers

TitleStatusHype
Point-GN: A Non-Parametric Network Using Gaussian Positional Encoding for Point Cloud ClassificationCode1
ViT-Lens: Initiating Omni-Modal Exploration through 3D InsightsCode1
Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud AnalysisCode2
ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D UnderstandingCode2
PointCLIP V2: Prompting CLIP and GPT for Powerful 3D Open-world LearningCode2
CLIP2Point: Transfer CLIP to Point Cloud Classification with Image-Depth Pre-trainingCode1
CALIP: Zero-Shot Enhancement of CLIP with Parameter-free AttentionCode1
PointCLIP: Point Cloud Understanding by CLIPCode1
Show:102550

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Point-GNAccuracy (%)86.4Unverified
2Point-NNAccuracy (%)64.9Unverified
3PointCLIP V2Accuracy (%)35.4Unverified
4CLIP2PointAccuracy (%)23.2Unverified
5CALIPAccuracy (%)16.9Unverified
6PointCLIPAccuracy (%)15.4Unverified