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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Most existing synthetic aperture radar (SAR) ship classification technologiesheavily rely on correctly labeled data, ignoring the discriminative features ofunlabeled SAR ship images. Even though researchers try to enrich CNN-basedfeatures by introducing traditional handcrafted features, existing methodseasily cause information redundancy and fail to capture the interaction betweenthem. To address these issues, we propose a novel dual-stream contrastivepredictive network (DCPNet), which consists of two asymmetric task designs andthe false negative sample elimination module. The first task is to constructpositive sample pairs, guiding the core encoder to learn more generalrepresentations. The second task is to encourage adaptive capture of thecorrespondence between deep features and handcrated features, achievingknowledge transfer within the model, and effectively improving the redundancycaused by the feature fusion. To increase the separability between clusters, wealso design a cluster-level tasks. The experimental results on OpenSARShip andFUSAR-Ship datasets demonstrate the improvement in classification accuracy ofsupervised models and confirm the capability of learning effectiverepresentations of DCPNet.

Dual-stream contrastive predictive network with joint handcrafted feature view for SAR ship classification