Procedurally generated water stains in synthetic training data reduce false alarms on real stains, and a sequential coreset variant of PatchCore makes high-resolution training memory-tractable.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Sequential PatchCore: Anomaly Detection for Surface Inspection using Synthetic Impurities
Procedurally generated water stains in synthetic training data reduce false alarms on real stains, and a sequential coreset variant of PatchCore makes high-resolution training memory-tractable.