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

Object detection in aerial imagery presents a significant challenge due tolarge scale variations among objects. This paper proposes an evolutionaryreinforcement learning agent, integrated within a coarse-to-fine objectdetection framework, to optimize the scale for more effective detection ofobjects in such images. Specifically, a set of patches potentially containingobjects are first generated. A set of rewards measuring the localizationaccuracy, the accuracy of predicted labels, and the scale consistency amongnearby patches are designed in the agent to guide the scale optimization. Theproposed scale-consistency reward ensures similar scales for neighboringobjects of the same category. Furthermore, a spatial-semantic attentionmechanism is designed to exploit the spatial semantic relations betweenpatches. The agent employs the proximal policy optimization strategy inconjunction with the evolutionary strategy, effectively utilizing both thecurrent patch status and historical experience embedded in the agent. Theproposed model is compared with state-of-the-art methods on two benchmarkdatasets for object detection on drone imagery. It significantly outperformsall the compared methods.

Scale Optimization Using Evolutionary Reinforcement Learning for Object Detection on Drone Imagery