Abstract: Hyperspectral anomaly detection (HAD) faces a significant challenge in separating scarce, small and subtle anomalous targets from complex backgrounds. For typical HAD approaches, background ...