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

Recently, the convolutional neural network has brought impressiveimprovements for object detection. However, detecting tiny objects inlarge-scale remote sensing images still remains challenging. First, the extremelarge input size makes the existing object detection solutions too slow forpractical use. Second, the massive and complex backgrounds cause serious falsealarms. Moreover, the ultratiny objects increase the difficulty of accuratedetection. To tackle these problems, we propose a unified and self-reinforcednetwork called remote sensing region-based convolutional neural network($\mathcal{R}^2$-CNN), composing of backbone Tiny-Net, intermediate globalattention block, and final classifier and detector. Tiny-Net is a lightweightresidual structure, which enables fast and powerful features extraction frominputs. Global attention block is built upon Tiny-Net to inhibit falsepositives. Classifier is then used to predict the existence of targets in eachpatch, and detector is followed to locate them accurately if available. Theclassifier and detector are mutually reinforced with end-to-end training, whichfurther speed up the process and avoid false alarms. Effectiveness of$\mathcal{R}^2$-CNN is validated on hundreds of GF-1 images and GF-2 imagesthat are 18 000 $\times$ 18 192 pixels, 2.0-m resolution, and 27 620 $\times$29 200 pixels, 0.8-m resolution, respectively. Specifically, we can process aGF-1 image in 29.4 s on Titian X just with single thread. According to ourknowledge, no previous solution can detect the tiny object on such huge remotesensing images gracefully. We believe that it is a significant step towardpractical real-time remote sensing systems.

R$^2$-CNN: Fast Tiny Object Detection in Large-Scale Remote Sensing Images