HyperAIHyperAI超神经

Command Palette

Search for a command to run...

算力平台

Point Cloud Registration On 3Dmatch Benchmark

评估指标

Feature Matching Recall

评测结果

各个模型在此基准测试上的表现结果

Paper Title
IMFNet98.6IMFNet: Interpretable Multimodal Fusion for Point Cloud Registration
MS-SVConv98.43D Point Cloud Registration with Multi-Scale Architecture and Unsupervised Transfer Learning
GeDi (no code published as of May 27 2021)97.9Learning general and distinctive 3D local deep descriptors for point cloud registration
SpinNet (no code published as of Dec 15 2020)97.6SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration
LMVD97.5End-to-End Learning Local Multi-view Descriptors for 3D Point Clouds
DIP + Point-TnT96.8Points to Patches: Enabling the Use of Self-Attention for 3D Shape Recognition
D3Feat-Pred95.8D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features
D3Feat-rand95.3D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features
DIP94.8Distinctive 3D local deep descriptors
3DSmoothNet94.7The Perfect Match: 3D Point Cloud Matching with Smoothed Densities
FCGF + RANSAC85Fully Convolutional Geometric Features
PPF-FoldNet71.8PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors
3DMatch + RANSAC66.83DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions
PPFNet62.3PPFNet: Global Context Aware Local Features for Robust 3D Point Matching
FPFH + RANSAC44.2Fast Point Feature Histograms (FPFH) for 3D Registration
0 of 15 row(s) selected.
Point Cloud Registration On 3Dmatch Benchmark | SOTA | HyperAI超神经