MoEnlight: Energy-efficient and self-Adaptive Low-light Video Stream Enhancement on Mobile Devices

Sicong Liu, Xiaochen Li, Zimu Zhou, Bin Guo, Yuan Xu, Zhiwen Yu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Camera-equipped devices and deep learning advancements have driven the development of intelligent mobile video apps. These apps require on-device processing of video streams for real-Time, high-quality services while addressing privacy and robustness. However, their performance is limited by low-light conditions and small-Aperture cameras in mobile platforms. Existing low-light video enhancement solutions are unsuitable due to complex models and lack of energy efficiency. We introduce MoEnlight, an energy-conscious system for enhancing low-light video on mobile devices. MoEnlight achieves real-Time enhancement with competitive quality, adapting to dynamic energy budgets. Our experiments demonstrate MoEnlight's superiority over state-of-The-Art solutions for enhancing low-light videos.

Original languageEnglish
Title of host publicationProceedings of ACM Turing Award Celebration Conference, CHINA 2023
PublisherAssociation for Computing Machinery, Inc
Pages19-20
Number of pages2
ISBN (Electronic)9798400702334
DOIs
StatePublished - 28 Jul 2023
Event2023 ACM Turing Award Celebration Conference, CHINA 2023 - Wuhan, China
Duration: 28 Jul 202330 Jul 2023

Publication series

NameProceedings of ACM Turing Award Celebration Conference, CHINA 2023

Conference

Conference2023 ACM Turing Award Celebration Conference, CHINA 2023
Country/TerritoryChina
CityWuhan
Period28/07/2330/07/23

Keywords

  • energy awareness
  • low light video enhancement
  • mobile devices

Fingerprint

Dive into the research topics of 'MoEnlight: Energy-efficient and self-Adaptive Low-light Video Stream Enhancement on Mobile Devices'. Together they form a unique fingerprint.

Cite this