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

Fine-grained object recognition concerns the identification of the type of anobject among a large number of closely related sub-categories. Multisource dataanalysis, that aims to leverage the complementary spectral, spatial, andstructural information embedded in different sources, is a promising directiontowards solving the fine-grained recognition problem that involves lowbetween-class variance, small training set sizes for rare classes, and classimbalance. However, the common assumption of co-registered sources may not holdat the pixel level for small objects of interest. We present a novelmethodology that aims to simultaneously learn the alignment of multisource dataand the classification model in a unified framework. The proposed methodinvolves a multisource region attention network that computes per-sourcefeature representations, assigns attention scores to candidate regions sampledaround the expected object locations by using these representations, andclassifies the objects by using an attention-driven multisource representationthat combines the feature representations and the attention scores from allsources. All components of the model are realized using deep neural networksand are learned in an end-to-end fashion. Experiments using RGB, multispectral,and LiDAR elevation data for classification of street trees showed that ourapproach achieved 64.2% and 47.3% accuracies for the 18-class and 40-classsettings, respectively, which correspond to 13% and 14.3% improvement relativeto the commonly used feature concatenation approach from multiple sources.

Multisource Region Attention Network for Fine-Grained Object Recognition in Remote Sensing Imagery