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

Offloading computationally heavy tasks from an unmanned aerial vehicle (UAV)to a remote server helps improve the battery life and can help reduce resourcerequirements. Deep learning based state-of-the-art computer vision tasks, suchas object segmentation and object detection, are computationally heavyalgorithms, requiring large memory and computing power. Many UAVs are using(pretrained) off-the-shelf versions of such algorithms. Offloading suchpower-hungry algorithms to a remote server could help UAVs save powersignificantly. However, deep learning based algorithms are susceptible tonoise, and a wireless communication system, by its nature, introduces noise tothe original signal. When the signal represents an image, noise affects theimage. There has not been much work studying the effect of the noise introducedby the communication system on pretrained deep networks. In this work, we firstanalyze how reliable it is to offload deep learning based computer vision tasks(including both object segmentation and detection) by focusing on the effect ofvarious parameters of a 5G wireless communication system on the transmittedimage and demonstrate how the introduced noise of the used 5G wirelesscommunication system reduces the performance of the offloaded deep learningtask. Then solutions are introduced to eliminate (or reduce) the negativeeffect of the noise. The proposed framework starts with introducing manyclassical techniques as alternative solutions first, and then introduces anovel deep learning based solution to denoise the given noisy input image. Theperformance of various denoising algorithms on offloading both objectsegmentation and object detection tasks are compared. Our proposed deeptransformer-based denoiser algorithm (NR-Net) yields the state-of-the-artresults on reducing the negative effect of the noise in our experiments.

Offloading Deep Learning Powered Vision Tasks from UAV to 5G Edge Server with Denoising