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

Underwater degraded images greatly challenge existing algorithms to detectobjects of interest. Recently, researchers attempt to adopt attentionmechanisms or composite connections for improving the feature representation ofdetectors. However, this solution does \textit{not} eliminate the impact ofdegradation on image content such as color and texture, achieving minimalimprovements. Another feasible solution for underwater object detection is todevelop sophisticated deep architectures in order to enhance image quality orfeatures. Nevertheless, the visually appealing output of these enhancementmodules do \textit{not} necessarily generate high accuracy for deep detectors.More recently, some multi-task learning methods jointly learn underwaterdetection and image enhancement, accessing promising improvements. Typically,these methods invoke huge architecture and expensive computations, renderinginefficient inference. Definitely, underwater object detection and imageenhancement are two interrelated tasks. Leveraging information coming from thetwo tasks can benefit each task. Based on these factual opinions, we propose abilevel optimization formulation for jointly learning underwater objectdetection and image enhancement, and then unroll to a dual perception network(DPNet) for the two tasks. DPNet with one shared module and two task subnetslearns from the two different tasks, seeking a shared representation. Theshared representation provides more structural details for image enhancementand rich content information for object detection. Finally, we derive acooperative training strategy to optimize parameters for DPNet. Extensiveexperiments on real-world and synthetic underwater datasets demonstrate thatour method outputs visually favoring images and higher detection accuracy.

Joint Perceptual Learning for Enhancement and Object Detection in Underwater Scenarios