Offloading the Computational Overhead of AI-application to the edge devices for face mask detection using Hybrid Computing Framework

Verma, Sumit (2022) Offloading the Computational Overhead of AI-application to the edge devices for face mask detection using Hybrid Computing Framework. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (3MB) | Preview
Preview	PDF (Configuration manual) Download (4MB) | Preview

Abstract

Cloud has become the important part of every business and organizational needs. Availability, scalability, Pay-per use model, easy connections with RESTful APIs are some of the popular features of cloud. Therefore, most of the business applications are highly depending on the cloud infrastructure to solve their business problems. On the other hand, Artificial intelligence technology is helping the businesses with automization of various applications and processes and these artificial intelligence-based applications utilizes the various deep learning and machine learning algorithms, which requires the large amount of storage and computing capabilities to train and run the applications. Due to the unavailability of large infrastructure, most of the AI-applications are being deployed to the cloud for model training and process the user requirements. The dependent application on cloud server requires high network bandwidth to establish the connection with cloud infrastructure, where latency is associated with each connection. The problem further arises when, the number of application users are much higher in number, which generates high amount of network traffic and computational overhead on the cloud server. In order to solve such challenges, this research is aimed to provide offloading-based hybrid computing architecture, where a small portion of code is installed on the edge device itself and dependency on the cloud is reduced. The study on the hybrid computing framework has been performed over the face mask detection application, which can detect the mask on the human face in the real time. This application has been deployed over both the cloud and hybrid computing based model and evaluated with different experiments and metrics.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Tamara Malone
Date Deposited:	16 Dec 2022 11:40
Last Modified:	07 Mar 2023 17:35
URI:	https://norma.ncirl.ie/id/eprint/5993

Actions (login required)

View Item