Keypoint-guided clustering with multi-prototype registration improves 3D point cloud anomaly detection on Real3D-AD, reaching 0.801 object-level and 0.861 point-level AUROC, though the margin over prior work is small and the fusion step is not scale-normalized.
R3d- ad: Reconstruction via diffusion for 3d anomaly detection,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
3DKeyAD: High-Resolution 3D Point Cloud Anomaly Detection via Keypoint-Guided Point Clustering
Keypoint-guided clustering with multi-prototype registration improves 3D point cloud anomaly detection on Real3D-AD, reaching 0.801 object-level and 0.861 point-level AUROC, though the margin over prior work is small and the fusion step is not scale-normalized.