Point Cloud Registration On 3Dlomatch 10 30

Metrics

Recall ( correspondence RMSE below 0.2)

Results

Performance results of various models on this benchmark

		Paper Title
GeoTransformer - P2PNet	74	Geometric Transformer for Fast and Robust Point Cloud Registration
NgeNet	71.9	Leveraging Inlier Correspondences Proportion for Point Cloud Registration
REGTR	64.8	REGTR: End-to-end Point Cloud Correspondences with Transformers
Predator-1k	62.5	PREDATOR: Registration of 3D Point Clouds with Low Overlap
Predator-5k	59.8	PREDATOR: Registration of 3D Point Clouds with Low Overlap
PCAM (reported in REGTR)	54.9	PCAM: Product of Cross-Attention Matrices for Rigid Registration of Point Clouds
DGR (reported in REGTR)	48.7	Deep Global Registration
FCGF (reported in PREDATOR)	40.1	Fully Convolutional Geometric Features
D3Feat (reported in PREDATOR)	37.2	D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features
3DSN (reported in PREDATOR)	33	The Perfect Match: 3D Point Cloud Matching with Smoothed Densities
Predator-NR	24	PREDATOR: Registration of 3D Point Clouds with Low Overlap
OMNet (reported in REGTR)	8.4	OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registration

0 of 12 row(s) selected.

Point Cloud Registration On 3Dlomatch 10 30 | SOTA | HyperAI