Preprint
Aug 2026
NFAD: Nuisance-Filtered Anomaly Detection Under Distribution Shift
Results show that explicitly suppressing such nuisance variation improves AD under distribution shift while preserving strong performance in standard settings, and NFAD remains competitive on standard benchmarks that do not explicitly evaluate distribution shift, including VisA, Real-IAD, and MVTec AD.
Dat Cao, Son T. Nghiem, Phan Nguyen et al.
· 0 citations