Multi-Prior Twin Least-Square Network for Anomaly Detection of Hyperspectral Imagery

Anomaly detection of hyperspectral imagery (HSI) identifies the very few samples that do not conform to an intricate background without priors. Despite the extensive success of hyperspectral interpretation techniques based on generative adversarial networks (GANs), applying trained GAN models to hyperspectral anomaly detection remains promising but challenging. Previous generative models can accurately learn the complex background distribution of HSI and typically convert the high-dimensional data back to the latent space to extract features to detect anomalies. However, both background modeling and feature-extraction methods can be improved to become ideal in terms of the modeling power and reconstruction consistency capability. In this work, we present a multi-prior-based network (MPN) to incorporate the well-trained GANs as effective priors to a general anomaly-detection task. In particular, we introduce multi-scale covariance maps (MCMs) of precise second-order statistics to construct multi-scale priors. The MCM strategy implicitly bridges the spectral- and spatial-specific information and fully represents multi-scale, enhanced information. Thus, we reliably and adaptively estimate the HSI label to alleviate the problem of insufficient priors. Moreover, the twin least-square loss is imposed to improve the generative ability and training stability in feature and image domains, as well as to overcome the gradient vanishing problem. Last but not least, the network, enforced with a new anomaly rejection loss, establishes a pure and discriminative background estimation.

Automated summary

In this work, a multi-prior-based network (MPN) is proposed to incorporate well-trained generative adversarial networks (GANs) as effective priors for hyperspectral anomaly detection. By introducing multi-scale covariance maps (MCMs) of precise second-order statistics, multi-scale priors are constructed to reliably and adaptively estimate the HSI label. The network is enhanced with twin least-square loss and a new anomaly rejection loss to improve generative ability and training stability, while establishing a pure and discriminative background estimation.