Citation: Al-qudah, R.; Khamayseh,
Y.; Aldwairi, M.; Khan, S. The Smart
in Smart Cities: A Framework for
Image Classification Using Deep
Learning. Sensors 2022, 22, 4390.
https://doi.org/10.3390/s22124390
Academic Editors: Panagiotis
Sarigiannidis, Thomas Lagkas,
Alexandros-Apostolos Boulogeorgos,
Vasileios Argyriou and Pantelis
Angelidis
Received: 4 May 2022
Accepted: 7 June 2022
Published: 10 June 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
The Smart in Smart Cities: A Framework for Image
Classification Using Deep Learning
Rabiah Al-qudah
1
, Yaser Khamayseh
2,3,
* , Monther Aldwairi
2,3
and Sarfraz Khan
4
1
Department of Computer Science and Software Engineering, Concordia University,
Montreal, QC H3G 1M8, Canada; r_alquda@encs.concordia.ca
2
College of Technological Innovation, Zayed University, Abu Dhabi 144534, United Arab Emirates;
monther.aldwairi@zu.ac.ae
3
Faculty of Computer and Information Technology, Jordan University of Science and Technology,
Irbid 22110, Jordan
4
ICT, Algonquin College, Ottawa, ON K2G 1V8, Canada; drkhansarfraz@gmail.com
* Correspondence: yaser.khamayseh@zu.ac.ae; Tel.: +971-2-599-3198
Abstract:
The need for a smart city is more pressing today due to the recent pandemic, lockouts,
climate changes, population growth, and limitations on availability/access to natural resources.
However, these challenges can be better faced with the utilization of new technologies. The zoning
design of smart cities can mitigate these challenges. It identifies the main components of a new
smart city and then proposes a general framework for designing a smart city that tackles these
elements. Then, we propose a technology-driven model to support this framework. A mapping
between the proposed general framework and the proposed technology model is then introduced. To
highlight the importance and usefulness of the proposed framework, we designed and implemented
a smart image handling system targeted at non-technical personnel. The high cost, security, and
inconvenience issues may limit the cities’ abilities to adopt such solutions. Therefore, this work also
proposes to design and implement a generalized image processing model using deep learning. The
proposed model accepts images from users, then performs self-tuning operations to select the best
deep network, and finally produces the required insights without any human intervention. This
helps in automating the decision-making process without the need for a specialized data scientist.
Keywords: smart city; deep learning; zoning; transfer learning; images; automation
1. Introduction
A smart city is a city that utilizes a set of the state-of-the-art technologies—information
and communication technologies, to provide better services to its citizens and users. This
general definition opens the door for many stakeholders to be involved in the design of
smart cities, such as computer scientists, software engineers, business managers, urban
developers, urban planners, and city officials. It is indeed hard to provide a unified
framework for smart cities, and it is even harder to measure the success of a smart city
in delivering its vision. The modern concept of smart cities evolved from the very first
initiatives toward digital cities in the 1990s [
1
]. Furthermore, evolved through the years to
a city that utilizes the emerging Internet of Things (IoT) technologies to achieve the smart
city strategic objectives [2].
The existence of suitable technological infrastructure is vital to the success of any
smart city project; hence, most projects were designed from a technological point of view
and were implemented to deliver technological solutions to overcome some of the current
limitations of traditional systems. However, considering the technological components
only in the design of a smart city framework is inadequate. Here, we identify some issues
that need to be handled in terms of both the people and the institution’s point of view:
Sensors 2022, 22, 4390. https://doi.org/10.3390/s22124390 https://www.mdpi.com/journal/sensors
