Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Edge analytics on resource constrained devices

Savitz, Sean, Perera, Charith and Rana, Omer 2021. Edge analytics on resource constrained devices. International Journal of Computational Science and Engineering
Item availability restricted.

[img] PDF - Accepted Post-Print Version
Restricted to Repository staff only

Download (2MB)


Video and image capture cameras have become an important type of sensor within the Internet of Things (IoT) sensing ecosystem. Camera sensors can measure our environment at high precision, providing the basis for detecting more complex phenomenon in comparison to other sensors e.g. temperature or humidity. This comes at a high computational cost on requirements of CPU, memory and storage resources, and requires consideration of various deployment constraints such as lighting and height of camera placement. Using benchmarks, this work evaluates object classification on resource-constrained devices, focusing on video feeds from IoT cameras. The models that have been used in this research include MobileNetV1, MobileNetV2 and Faster R-CNN that can be combined with regression models for precise object localisation. We compare the models by using their accuracy for classifying objects and the demand they impose on the computational resources of a Raspberry Pi. Various IoT deployments are investigated by comparing the probability scores of classifying chosen objects using different camera placements. We conclude that the Faster R-CNN model that is configured with the InceptionV2 regression model has the highest accuracy. However, this is at the cost of additional computational resources. We found that the best model to use for object detection functionality on the Raspberry Pi is the MobileNetV2 model paired with the SSDLite regression model. This results in the highest accuracy and probability score for object classification, in comparison to other mobile-friendly models considered in this work, whilst using the least amount of computational resources.

Item Type: Article
Status: In Press
Schools: Computer Science & Informatics
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Publisher: Inderscience
ISSN: 1742-7185
Date of First Compliant Deposit: 15 March 2021
Date of Acceptance: 2 March 2021
Last Modified: 15 Mar 2021 16:30

Actions (repository staff only)

Edit Item Edit Item


Downloads per month over past year

View more statistics