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

In recent years, convolutional networks have demonstrated unprecedentedperformance in the image restoration task of super-resolution (SR). SR entailsthe upscaling of a single low-resolution image in order to meetapplication-specific image quality demands and plays a key role in mobiledevices. To comply with privacy regulations and reduce the overhead of cloudcomputing, executing SR models locally on-device constitutes a key alternativeapproach. Nevertheless, the excessive compute and memory requirements of SRworkloads pose a challenge in mapping SR networks on resource-constrainedmobile platforms. This work presents MobiSR, a novel framework for performingefficient super-resolution on-device. Given a target mobile platform, theproposed framework considers popular model compression techniques and traversesthe design space to reach the highest performing trade-off between imagequality and processing speed. At run time, a novel scheduler dispatchesincoming image patches to the appropriate model-engine pair based on thepatch's estimated upscaling difficulty in order to meet the required imagequality with minimum processing latency. Quantitative evaluation shows that theproposed framework yields on-device SR designs that achieve an average speedupof 2.13x over highly-optimized parallel difficulty-unaware mappings and 4.79xover highly-optimized single compute engine implementations.

MobiSR: Efficient On-Device Super-Resolution through Heterogeneous Mobile Processors