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

Object detection in optical remote sensing images is an important andchallenging task. In recent years, the methods based on convolutional neuralnetworks have made good progress. However, due to the large variation in objectscale, aspect ratio, and arbitrary orientation, the detection performance isdifficult to be further improved. In this paper, we discuss the role ofdiscriminative features in object detection, and then propose a CriticalFeature Capturing Network (CFC-Net) to improve detection accuracy from threeaspects: building powerful feature representation, refining preset anchors, andoptimizing label assignment. Specifically, we first decouple the classificationand regression features, and then construct robust critical features adapted tothe respective tasks through the Polarization Attention Module (PAM). With theextracted discriminative regression features, the Rotation Anchor RefinementModule (R-ARM) performs localization refinement on preset horizontal anchors toobtain superior rotation anchors. Next, the Dynamic Anchor Learning (DAL)strategy is given to adaptively select high-quality anchors based on theirability to capture critical features. The proposed framework creates morepowerful semantic representations for objects in remote sensing images andachieves high-performance real-time object detection. Experimental results onthree remote sensing datasets including HRSC2016, DOTA, and UCAS-AOD show thatour method achieves superior detection performance compared with manystate-of-the-art approaches. Code and models are available athttps://github.com/ming71/CFC-Net.

CFC-Net: A Critical Feature Capturing Network for Arbitrary-Oriented Object Detection in Remote-Sensing Images