Page 149 - 2023-Vol19-Issue2

Received: 1 May 2023 | Revised: 9 July 2023 | Accepted: 19 July 2023

DOI: 10.37917/ijeee.19.2.17 Vol. 19 | Issue 2 | December 2023

Open Access

Iraqi Journal for Electrical and Electronic Engineering

Original Article

Content-Based Image Retrieval using Hard Voting

Ensemble Method of Inception, Xception, and Mobilenet

Architectures

Meqdam A. Mohammed*, Zakariya A. Oraibi, Mohammed Abdulridha Hussain
Department of Computer Science, College of Education for Pure Sciences, University of Basrah, Basrah 61004, Iraq

Correspondance
* Meqdam A. Mohammed
Baghdad, Iraq
Email: mkdaam@gmail.com

Abstract
Advancements in internet accessibility and the affordability of digital picture sensors have led to the proliferation
of extensive image databases utilized across a multitude of applications. Addressing the semantic gap between low-
level attributes and human visual perception has become pivotal in refining Content Based Image Retrieval (CBIR)
methodologies, especially within this context. As this field is intensely researched, numerous efficient algorithms for
CBIR systems have surfaced, precipitating significant progress in the artificial intelligence field. In this study, we
propose employing a hard voting ensemble approach on features derived from three robust deep learning architectures:
Inception, Exception, and Mobilenet. This is aimed at bridging the divide between low-level image features and human
visual perception. The Euclidean method is adopted to determine the similarity metric between the query image and the
features database. The outcome was a noticeable improvement in image retrieval accuracy. We applied our approach
to a practical dataset named CBIR 50, which encompasses categories such as mobile phones, cars, cameras, and cats.
The effectiveness of our method was thereby validated. Our approach outshone existing CBIR algorithms with superior
accuracy (ACC), precision (PREC), recall (REC), and F1-score (F1-S), proving to be a noteworthy addition to the field
of CBIR. Our proposed methodology could be potentially extended to various other sectors, including medical imaging
and surveillance systems, where image retrieval accuracy is of paramount importance.

Keywords
CBIR, Ensemble Learning, Deep Learning, Classification, Hard voting.

I. INTRODUCTION Visions (CVs) and Artificial Intelligence (AI) domains [1].
The two primary techniques or elements of a CBIR system are
The increasing usage of digital devices and developments in picture representation for picture classification and similarity
internet technology have made it simple and convenient to take measure for search query. It is assumed that feature vectors
pictures of any desired thing. As a consequence, a significant or image representations will be discriminative in order to
amount of photos are produced every day, which may be used discriminate between pictures.
to improve processing information efficiency and make daily
life more logical and comfortable. The use of Content-Based Moreover, it is anticipated that it will be invariant to spe-
Image Retrieval (CBIR) methods is one way to make use of cific modifications. The similarity between two photos should
these photos. These methods enable the use of an input image reflect the semantic importance based on how the images are
of the desired item or content to get pertinent photographs represented. These two interconnected components play a key
from a database. CBIR is still a useful tool for image retrieval role in retrieval performance and the existing CBIR algorithms
and processing despite being widely used in many Vomputer may be grouped based on how well they contribute to these

This is an open-access article under the terms of the Creative Commons Attribution License,
which permits use, distribution, and reproduction in any medium, provided the original work is properly cited.
©2023 The Authors.
Published by Iraqi Journal for Electrical and Electronic Engineering | College of Engineering, University of Basrah.

https://doi.org/10.37917/ijeee.19.2.17 |https://www.ijeee.edu.iq 145