Compact Neural Architecture Balances Visual Anomaly Detection Accuracy and Hardware Costs
A compact neural network architecture combines local receptive fields, lateral inhibition, and feature-space reconstruction to deliver competitive industrial anomaly detection ...

