Quality-of-Things Based Machine Learning for the MIoT Applications

Shaymaa Al-Juboori, Ali H. Husseen Al-Nuaimi, Amulya Karaadi, Is Haka Mkwawa, Jianwu Zhang, Lingfen Sun

Research output: Contribution to conferenceConference paper (not formally published)peer-review

10 Downloads (Pure)

Abstract

Enhancing the Quality of Things (QoT) is urgently needed given the rapid evolution of the Multimedia Internet of Things (MIoT). One of the challenges with MIoT is Acceptable QoT. Achieving AQoT can optimize bandwidth and storage at a level that will satisfy MIoT application’s minimal requirements. Intelligent systems using Machine Learning (ML) techniques can improve the performance of MIoT applications by keeping the minimum requirements of the resources which are necessary to maintain AQoT. The aim of this study is to develop a MIoT system based on ML to provide high performance with an acceptable QoT. The Gaussian-Naive Bayes, Fine KNN, and AdaBoost ML algorithms were investigated against different video sequences with varying bitrates and network conditions. The results based on face recognition for a Ring Video Doorbell scenario showed that ML could be used on MIoT applications to achieve AQoT and significantly reduce bandwidth and storage usage.
Original languageEnglish
DOIs
Publication statusPublished - Jun 2023
Event2023 IEEE International Conference on Acoustics, Speech and Signal Processing - Greece, Greek island of Rhodes
Duration: 4 Jun 202310 Jun 2023
https://2023.ieeeicassp.org/

Conference

Conference2023 IEEE International Conference on Acoustics, Speech and Signal Processing
Abbreviated titleICASSP 2023
CityGreek island of Rhodes
Period4/06/2310/06/23
Internet address

Fingerprint

Dive into the research topics of 'Quality-of-Things Based Machine Learning for the MIoT Applications'. Together they form a unique fingerprint.

Cite this