HyperAI超神经
首页资讯最新论文教程数据集百科SOTALLM 模型天梯GPU 天梯顶会
全站搜索
关于
HyperAI超神经

Zero Shot Transfer 3D Point Cloud

评估指标

Accuracy (%)

评测结果

各个模型在此基准测试上的表现结果

模型名称
Accuracy (%)
Paper TitleRepository
PointCLIP V264.22PointCLIP V2: Prompting CLIP and GPT for Powerful 3D Open-world Learning
OpenShape-SparseConv (+dlign)85.0OpenDlign: Open-World Point Cloud Understanding with Depth-Aligned Images
ReCon61.7Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
CLIP2Point49.38CLIP2Point: Transfer CLIP to Point Cloud Classification with Image-Depth Pre-training
PointCLIP20.18PointCLIP: Point Cloud Understanding by CLIP
OpenDlign82.6OpenDlign: Open-World Point Cloud Understanding with Depth-Aligned Images
ULIP + PointMLP61.5ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understanding
ReCon++87.3ShapeLLM: Universal 3D Object Understanding for Embodied Interaction
OpenShape-PointBERT (+dlign)85.4OpenDlign: Open-World Point Cloud Understanding with Depth-Aligned Images
ViT-Lens87.6ViT-Lens: Initiating Omni-Modal Exploration through 3D Insights
OpenShape-SparseConv83.4OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understanding
MixCon3D-PointBERT86.8Sculpting Holistic 3D Representation in Contrastive Language-Image-3D Pre-training
OpenShape-PointBERT85.3OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understanding
Uni3D88.2Uni3D: Exploring Unified 3D Representation at Scale
ULIP + PointBERT60.4ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understanding
TAMM-PointBERT (+dlign)86.2OpenDlign: Open-World Point Cloud Understanding with Depth-Aligned Images
0 of 16 row(s) selected.