Abstract: Hyperspectral anomaly detection (HAD) can be formulated as a binary classification problem, where the accurate separation of anomalies from the background relies on the precise features ...