Geometry-aware 3D point cloud anomaly detection via normal-augmented memory banks
A unified framework to address gaps in memory-bank-based 3D point cloud anomaly detection via two integrated components: Normal-Aware Registration Features (NARF), which augments the RANSAC--ICP pipeline to extract per-point orientations with negligible overhead, and Geometry-Aware Coreset Sampling (GACS), which priori...