English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Hyperspectral images (HSIs) record the electromagnetic energy from the visible to infrared bands, which can obtain rich spectral and spatial information. HSI is a three-dimensional volume, which combines one-dimensional spectral feature and two-dimensional spatial information. Hyperspectral anomaly detection aims at detecting the objects where the spectral or spatial features are significantly different from the neighbouring background when the targets or background priors are unknown, which is a special kind of anomaly detection. In this paper, we propose a hyperspectral anomaly detector based on vision-transformer. We cast anomaly detection as an image inpainting problem, and the background is reconstructed by the inpainting-transformer model, whereas the anomaly is detected by a morphological filter. Since the background is large, whereas the anomaly area is relatively smaller, we randomly mask out some areas and train the network to recover these areas based on the background context, and the network mainly learns the background feature from the surroundings, thus the background area is reconstructed at first. Then, we remove the background from the input image and use the morphological filter to extract small areas, which enhances the detection results. Finally, we use an adversarial strategy to suppress the reconstruction of anomalies and refine the results based on suppressed images. Extensive experiments on two HSI dataset with 15 images demonstrate that our method outperforms state-of-the-art competitors.

Hyperspectral anomaly detection with vision transformer and adversarial refinement