Received 3 December 2022, accepted 25 December 2022, date of publication 5 January 2023, date of current version 13 January 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3234279

Diagnosis of Malaria Using Double Hidden Layer

Extreme Learning Machine Algorithm With CNN
Feature Extraction and Parasite Inflator
MD. OMAER FARUQ GONI 1,∗ , MD. NAZRUL ISLAM MONDAL2,∗ , (Senior Member, IEEE),
S. M. RIAZUL ISLAM 3,∗ , (Member, IEEE), MD. NAHIDUZZAMAN 1,5,∗ , MD. ROBIUL ISLAM 1,

MD. SHAMIM ANOWER 4 , (Member, IEEE), AND

KYUNG-SUP KWAK 6 , (Life Senior Member, IEEE)
1 Department of Electrical and Computer Engineering, Rajshahi University of Engineering and Technology, Rajshahi 6204, Bangladesh
2 Department of Computer Science and Engineering, Rajshahi University of Engineering and Technology, Rajshahi 6204, Bangladesh
3 Department of Computer Science, University of Huddersfield, HD1 3DH Huddersfield, U.K.
4 Department of Electrical and Electronic Engineering, Rajshahi University of Engineering and Technology, Rajshahi 6204, Bangladesh
5 Department of Electrical Engineering, Qatar University, Doha, Qatar
6 Department of Information and Communication Engineering, Inha University, Incheon 22212, South Korea

Corresponding authors: Md. Omaer Faruq Goni ([email protected]) and Kyung-Sup Kwak ([email protected])
This work was supported by the National Research Foundation of Korea-Grant funded by the Korean Government (Ministry of Science
and ICT) under Grant NRF-2020R1A2B5B02002478.
*Md. Omaer Faruq Goni, Md. Nazrul Islam Mondal, S. M. Riazul Islam, and Md. Nahiduzzaman contributed equally to this work.

ABSTRACT Malaria, a life-threatening disease worldwide, can be diagnosed using antigen tests and
microscopy tests. However, both of them are erroneous and time-consuming. Therefore, a trustworthy and
fast early malaria prognosis infrastructure is required. In this age of machine learning (ML), there are several
ML-based methods to do the task. This paper proposes an unorthodox method for malaria prognosis based on
an extreme learning machine (ELM) algorithm. In this regard, Convolutional Neural Networks (CNN), ELM,
and double hidden layer (DELM) have been used as classifiers. A CNN model has been used as a feature
extractor and also as a classifier to perform a comparative study. The derived features have been used to train
ELM and DELM. Two versions of the malaria image dataset have been used: one is the original dataset, and
the other is a modified dataset where ambiguous samples have been removed. The parasite inflator acts
as the shape increaser of the small, darker malaria parasites in the RBC images in order to detect malaria
easily. CNN-DELM has achieved a sanguine result on every performance standard compared to CNN and
CNN-ELM. The proposed CNN-DELM method has achieved 97.79% and 99.66% accuracy for the original
version and the modified version, respectively. Hence, the proposed CNN-DELM model has also produced
either comparable or better results when compared to other methods proposed in the literature, showing its
robustness in detecting malaria.

INDEX TERMS Convolutional neural network (CNN), double hidden layer extreme learning machine
(DELM), malaria, extreme learning machine (ELM).

I. INTRODUCTION yearly [1]. South-East Asia, the Eastern Mediterranean, the

In 2015, the World Health Organization (WHO) reported Western Pacific, and the Americas have all been recognized
that around 438 thousand people died due to malaria para- as high-risk regions by the WHO. Malaria is a dangerous
sites, and 620 thousand people died in 2017, and approxi- and deadly disease that is triggered from the bite of female
mately 300-500 million people were affected by this disease anopheles mosquitoes that host plasmodium parasites. There
are around 400 types of anopheles; among them, 30 types
The associate editor coordinating the review of this manuscript and mainly act as parasite carrieres. To be a parasite carrier, the
approving it for publication was Yu-Huei Cheng . female anopheles mosquito has to bite a malaria-infected

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME 11, 2023 4117
 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

person. The life cycle of the anopheles mosquito consists of of different approaches are presented in Section IV, and the
four stages: egg, larva, pupa, and adult. To bring the eggs into outcomes compared to the findings of other recent studies.
the adult stage, they have to feed on human blood. In this feed- The key conclusions are presented at the end in section V.
ing process, malaria spreads all around. There are five Plas-
modium parasite species, including Plasmodium falciparum, II. RELATED WORK
P. vivax, P. ovale, P. malariae, and P. knowlesi. Among them, Malaria, as a life-threatening disease, has attracted the curios-
P. falciparum and P. vivax are the most dangerous. It takes ity of researchers all around the world. Malaria was formerly
10 to 15 days or even more to remain in hibernation after a mainly diagnosed in the laboratory, demanding a massive
carrier mosquito bite. After the hibernation period, it seizures lot of human knowledge. Automatic systems based on ML,
the red blood cells (RBC) and reduces the number of RBC, DL have been utilized by most researchers to detect malaria
revealing different indications of malaria, such as fever, chills, from RBC images nowadays. In this section, some of these
nausea, and vomiting [2]. Malaria should be diagnosed as recent and well-known studies have been described.
soon as possible because it can rapidly turn into a severe Rajaraman et al. developed a customized deep CNN
stage and is life-threatening. It cannot be passed from one model to extract features from parasitized and uninfected
person to another, however malaria can be transmitted from cells images [9]. They have utilized various transfer learn-
mother to fetus, contracted through blood transfusions or ing (TL) algorithms, for instance, AlexNet [10], VGG-16
sharing injections [3], [4]. This disease can be spread in hot, [11], ResNet-50 [12], Xception [13], etc., for the detection
humid climates near natural water sources where Anopheles of malaria and achieved a high accuracy of 0.959 using
mosquitoes transmit deadly diseases [5]. VGG-16. Masud et al. proposed leveraging deep CNN for
In general, there are two traditional ways to diagnose real-time detection of malaria from RBC images [14]. They
malaria where; one is thin blood cell microscopy, however it developed a custom CNN using cyclical stochastic gradient
is a prolonged process; usually, a microscopist must manually descent (SGD) as an optimizer and achieved an accuracy
identify a minimum of 5000 cells to certify the condition. of 97.30%. Rosado et al. have used SVM to look at how
The second is an antigen diagnostic examination; despite to identify malaria parasites and white blood cells using
being significantly faster than the former, antigen-based rapid smartphones [15].
diagnostic tests are error-prone and costly. In underdevel- Maqsood et al. utilized various TL algorithms for the
oped countries, patients cannot be admitted to prompt care, detection of malaria from RBC images [16]. Further, they
or antigen-based expedited therapy diagnostic studies cannot developed a customized deep CNN model that outperforms
be reimbursed. Hence, for this reason, it is crucial to formu- the other TL models. Before training their customized model,
late a system that diagnoses malaria quickly and accurately, they utilized bilateral filtering and image augmentation tech-
and the system must be cost-effective. niques to emphasize features of RBC. Sriporn et al. utilized
In the last decade, researchers have proposed several the Xception TL model with various types of activation
machine learning (ML) and deep learning (DL) based auto- functions and optimizers for the detection of malaria [17].
matic diagnosis systems to detect various life-threatening They achieved high accuracy of 99.28% while combined
diseases [6], [7], [8]. In this study, the advantages of both ML Nadam optimizer with Mish activation function. Jain et al.
and DL have been merged to detect malaria from RBC images proposed a low-cost, simple CNN with no preprocessing
more efficiently. cell images and achieved an accuracy of 97% [18]. Dong
Some RBC samples in the original dataset were misla- et al. proposed DL methods for automatic identification of
beled. This problem has been resolved in the updated dataset. malaria-infected cells [19]. They utilized three transfer learn-
In this study, both datasets have been used. Some prepro- ing models including the LeNet, AlexNet, and GoogLeNet,
cessing processes have been utilized, which are referred to and achieved an accuracy of over 95%. Yang et al. customized
as ‘‘parasitic inflators’’ in the group. As a feature extractor, a CNN for the classification of malaria parasites in thick
a lightweight CNN has been used, while DELM has been used blood smear images [20]. To select parasite candidates, they
as the final classifier. used an intensity-based Iterative Global Minimum Screening
The main contributions of this paper are as follows: (IGMS) on a thick smear picture and achieved an accuracy of
93.46% ± 0.32% and AUC of 98.39% ± 0.18%. Shah et al.
• Two dataset: the original and the modified malaria
proposed a deep CNN for the detection of malaria from RBC
dataset, has been used in this study
images and achieved an accuracy of 95% [21].
• Small darker parasite spot has been increased in shape
Khan et al. derived the aggregated features from RC images
using proposed ‘‘Parasite Inflator’’ to detect malaria
using the aggregated laplacian coefficient [22]. After that,
easily
they utilized a random forest classifier to classify malaria
• The proposed CNN-DELM model shows the best per-
from these derived features and achieved a recall of 86%.
formance for both original and updated dataset.
Olugboja et al. utilized a support vector machine (SVM) and
The rest of the paper is organized as section II that repre- CNN to detect malaria from a private dataset of 2565 RCB
sents the recent works on malaria classification. Section III images collected from the University of Alabama at Birm-
describes different steps of the proposed method. The results ingham [23]. They achieved an accuracy of 95% and 91.66%

4118 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 1. Proposed DELM framework for detection of malaria.

using CNN and SVM, respectively. Fuhad et al. derived proposed a novel deep belief network (DBN) for the classi-
features from RBC cells images using CNN for automatic fication of malaria from RBC images [31]. For the classifi-
diagnosis of malaria [24]. They also performed various pre- cation of 4100 images of peripheral blood smears into the
processing techniques, for instance, knowledge distillation, parasite or non-parasitic class, they utilized a trained model
data augmentation, etc. For the classification of malaria, they based on a DBN and achieved f1-score and sensitivity of
used SVM of k-nearest neighbors (KNN) and achieved an 89.66%, and 97.60% respectively.
accuracy of 99.23%. Anggraini et al. introduced a system that
uses image segmentation techniques to separate blood cells III. PROPOSED DELM ARCHITECTURE AND
from their background [25]. They performed segmentation CONSTITUENTS
by global thresholding to retrieve erythrocyte and other blood For the previous few decades, researchers have concentrated
cell components in each image. Gopakumar et al. focused on on developing various computer-aided systems that can help
a stack of images and developed custom CNN [26]. The cell detect many life-threatening disorders from the image of a
counting problem was recast as a segmentation problem, and medical exam. Many diseases such as Malaria, heart dis-
a two-level segmentation technique was proposed. ease, breast cancer, brain tumor and some other disorders,
Liang et al. proposed an ML technique based on a CNN all demand early detection to save lives. It is critical to
to automatically categorize single cells in thin films [27]. create a system that identifies and addresses these diseases
Based on 27,578 singles, ten-fold cross-validation was per- early on to save money and lives. This research introduces
formed and achieved an average accuracy of 97.37% using a new architecture of diagnosing malaria parasites. Kaggle
a new 16-layer CNN model on cell images. Tomari et al. has provided the RBC images. For pre-processing of the cell
utilized the global threshold approach on a green channel images, this study used recognized image processing tech-
color image to extract RBCs from the background [28]. Then, niques such as morphological processing and other processes,
a morphological filter and related component labeling were which are explained later. Once pre-processing has been
used, noise and holes in the RBCs were removed. Following done, a lightweight CNN has been designed to extract and
that, the geometrical features of the RBCs are used to extract find the most informative attributes. Additionally, a newly
information from them. They utilized an Artificial Neural structured DELM method has been presented for identifying
Network (ANN) to detect malaria from this derived informa- malaria-infected cells. Figure1 shows the approach under
tions. Diaz et al. utilized an SVM to identify preprocessed consideration.
blood smear images to detect infectious erythrocytes [29].
They used a dataset of 450 malaria images, and their model A. DESCRIPTION OF DATASET
performed well in terms of specificity and sensitivity. The malaria dataset has been collected from the Kaggel [32].
Pattanaik et al. proposed a computer-aided diagnosis It contains 27,558 images of RBC. The dataset contain equal
(CAD) model for the detection of malaria from cell amount of malaria infected and uninfected samples hence it
images [30]. They used an artificial neural network with a is a balance dataset.
functional connection and sparse stacking is used to pre-train When analysis of the dataset has been perform, some of
the parameters of the system and achieved a classification the samples have been arisen confusion. It shows the prob-
accuracy of 89.10% and sensitivity of 93.90%. Bibin et al. ability of miss-labeling of some samples. Fuhad [24] has

VOLUME 11, 2023 4119

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

also got same problem and to solve this problem, the authors Algorithm 1 Parasite Inflator
have been consulted with medical expert. They removed Input: RBC cell image
647 falsely labeled parasitized data which was considered as Output: Preprocessed RBC cell image
false parasitized and 750 falsely labeled uninfected images 1) Define a 4 × 4 rectangular kernel
which were named as false uninfected. The samples have 2) img : Read the RBC cell image
been labelled correctly. They have shared the updated sam- 3) img_erode : Perform erosion 15 times
ples through google drive [33]. 4) img_resized : Resize the image into 32 × 32
The updated dataset contains 13,132 infected and 13,029 5) img_norm : Convert the pixel values between 0 to 1
uninfected samples. This study has been carried out using
both the original and updated dataset. Figure 2 shows some
samples regarding labeling issue.
set [34]. It will help handle more data and get better data
B. PRE-PROCESSING management. Because of the intricate imaging attributes for
Image pre-processing is a crucial step for this type of study cell images, a new CNN model for finding 512 of the most
because model outcome highly depends on pre-processing significant aspects has been designed for malaria identifica-
techniques applied. It makes the learning process smooth. tion. The Figure shows the proposed a shallow CNN model.
The proposed CNN model utilizes three convolutional lay-
1) PARASITE INFLATOR ers and two fully connected layers. The convolutional layers
To detect malaria, thin or thick blood seamer is stained with use batch normalization and max-pooling after each layer.
chemicals and then a microscope is used to recognize the par- The inputs of the layers are normalized to speed up running
asite. Due to the chemical effect, the parasite can be identified and increase stability [35]. A pooling layer has been put
as a darker spot in the blood seamer. after each convolutional layer. The researchers used Max-
In this study, any small darker spot that represents a para- pooling with 2 × 2 filters, which enable the extraction of
site has been increased in shape so that it is easy to classify. the most significant parts of images via the capability to
As the background of the malaria images is black, at first obtain the greatest value in each cluster at the convolutional
the background has been converted into white. After that, layers [36], [37], [38]. Figure 4 shows the proposed CNN
the darker parasite spots have been inflated using erosion. for features extraction. First, the ‘‘SAME’’ padding has been
To perform erosion, a 4 × 4 sized rectangular kernel has included in the first convolution layer as the output results
been used. To inflate enough, the erosion has been performed from applying filters to all image tuples. Border components
15 times. Figure 3 represents some samples before and after have been examined because of the features that can be
preprocessing. After preprocessing the small sized parasite included. No extra padding has been used to compute the
are increased in size that is easily recognizable. border elements. On the contrary, the ’VALID’ padding
ignored the border components. To avoid gradient vanishing,
2) RESIZING ReLU has been utilized as an activation function [18]. After
The malaria dataset contains images of different sizes. Same the final dense layer, a sigmoid has been introduced.
size of images is required to conduct the analysis. Due, Dropout regularization has been used to overcome the
to make the analysis faster the images have been resized overfitting problem. Throughout the training phase of a neu-
32 × 32 pixels in size. Images have been resized after apply- ral network, randomly selected nodes are discarded during
ing the parasite inflator. the weight update phases [39]. The Adam optimizer was used
because it is exact for CNNs, improves training on vast vol-
3) NORMALIZATION umes of data, and for calculating the loss binary cross-entropy
Pixel intensity of an image varies from 0 to 255. Each pixel has been used [40]. Using the developed model, the learning
value acts as a feature in the CNN model. To convert all the rate is 0.001, and the model was trained for 100 epochs, with
pixel values into the range of 0 to 1 for reducing the learning a batch size of 1024. And finally, 512 prominent features have
complexity, normalization is used. It can be done by dividing been derived from the last dense layer. Here is a summary of
image pixel values by 255. the lightweight CNN model shown in Table 1.
The preprocessing can be represented in this way:
D. EXTREME LEARNING MACHINE
C. FEATURE EXTRACTION USING CNN To reduce the training time complexity caused by the repeated
The important component of the classification issue is feature model parameter tuning process, ELM has been developed by
extraction since the success of a model depends on how Huang et al. [41]. It is a neural network (NN) which is feed-
accurately it can find the most relevant elements in the RBC forwarding consisting of an input layer, single hidden layer
images. It needs to identify the important characteristics dis- and an output layer. It can attain the lowest training error. For
tinguished between the two categories to improve the model’s a regular NN, network parameters such as weight and bias are
classification ability. A feature extraction approach is used chosen arbitrarily and the back-propagation (BP) algorithm
to create a condensed, informative, lower-dimensional data is used to optimize the parameters. In contrast to regular NN,

4120 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 2. Label analysis.

TABLE 1. Brief architectural overview of proposed lightweight CNN for architecture is simple and does not have iterative parameter
feature extractor.
tuning that makes the training process faster and achieves
adequate performance in disease classification [42], [43].
Figure 5 shows the proposed DELM model for malaria clas-
sification. In ELM, 500 hidden nodes have been used. In this
study, DELM has also been used. It has been consisted of
500 hidden nodes in each layer. There is no vanishing gradient
issue with ELM. However, the ReLu activation function was
employed with hidden nodes to provide non-linearity in this
case. The number of hidden nodes and the ReLu activation
function have been chosen through trial and error method.
ELM can be represented using following steps:

Algorithm 2 Extreme Learning Machine

1) Randomly generates the input weight W(m,N ) and bias
B(1,N ) matrix.
2) Determine the output H(n,N ) of the hidden layer.
H(n,N ) = G(X(n,m) · W(m,N ) + B(1,N ) )
3) Determine the output weight matrix β(N ,t) β(N ,t) =
†
H(N ,n) · T(n,t)
4) Make prediction using β(N ,t)
though the weights between input and hidden layer are chosen
arbitrarily, no optimization process is used. And the weight DELM has been proposed by Ding [44]. DELM can imple-
between the hidden and output layer is calculated analytically mented in the similar manner. The required step are given
using the Moore-Penrose Pseudoinverse method. The ELM bellow [45]:

VOLUME 11, 2023 4121

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 3. Pre-processed images.

FIGURE 4. Proposed CNN for features extraction.

IV. RESULT AND DISCUSSION Accuracy is defined as the percentage of accurately

A. EVALUATION MATRICS detected cases among all cases. It demonstrates how good the
For determining the effectiveness of the models, several eval- classification system is at detecting patterns [46], [47].
uation standards have been addressed for instance accuracy, (TP + TN )
Accuracy = (1)
precision, recall, f1-score, and receiver operating character- (TP + TN + FP + FN )
istic (ROC) curve. All of these criteria have been calculated where true positive, true negative, false positive, and false
using a confusion matrix. negative have been represented as TP, TN, FP, and FN. In this

4122 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 6. Confusion matrix of CNN with original images.

TABLE 2. Results of CNN with original images.

FIGURE 5. Proposed DELM for classification.

Algorithm 3 Double Hidden Layer Extreme Learning

Machine
1) Randomly generates the input weight W(m,N ) and bias
B(1,N ) matrix.
2) Determine the output H1(n,N ) of the hidden layer.
H1(n,N ) = G(X(n,m) · W1(m,N ) + B1(1,N ) )
3) Determine weight between first hidden layer and out-
†
put layer. Temp_β(N ,t) = H1(N ,n) · T(n,t)
4) Determine the expected output of 2nd hidden layer.
†
H2 _Exp(n,N ) = T(n,t) · Temp_β(N ,t)
5) Calculate the weight between 1st and 2nd hidden
†
layer. W2 = H1(N ,n) · G† (H2 _Exp(n,N ) )
6) Calculate the actual output of the 2nd hidden layer
H2(n,N ) = G(H1(n,N ) · W2(m,N ) + B2(1,N ) ) FIGURE 7. Confusion matrix for CNN-ELM with original images.
7) Determine the output weight matrix β(N ,t) β(N ,t) =
†
H2(N ,n) · T(n,t) A PC with a 64-bit Windows 10 operating system has been
8) Make prediction using all the wights. used.

study, precision has been measured as the percentage of C. ORIGINAL DATASET

patients who properly recognized having malaria out of all This section discusses the experiments that have been con-
those who had the disease [48]. ducted and the results obtained through the use of various
performance measurements. First, the RBC images have been
TP
Precision = (2) collected from Kaggle and pre-processed. Then developed
(TP + FP) a shallow CNN for extracting the most informative features
The recall has been measured as the percentage of malaria from these processed images and also used for the classifica-
patients must be correctly identified as malaria out of all tion of malaria. Finally, ELM and DELM have been proposed
patients [48]. to detect malaria more accurately and efficiently than CNN.
TP The algorithms have been trained using 24802 processed
Recall = (3) images where 12387 images from normal, 12415 images
(TP + FN )
are from infected patients with malaria. To evaluate the
F1-score has been calculated using the following equa- performance and effectiveness of the models, 2756 data has
tion [49] been used for testing the models where 1392 from uninfected
2 ∗ (Precision ∗ Recall) and 1364 cell images from infected patients.
F1 − score = (4)
(Precision + Recall) A confusion matrix (CM) has been used to calcu-
late the accuracy, precision, recall, f1-score, and AUC
B. ENVIRONMENTAL SETUP shown to examine the reliability of the proposed models.
An online platform kaggle used for running the python code Figure 6, 7 and 8 demonstrated the CM for CNN, CNN-ELM
with TPUs. Keras has been used to implement the algorithms, and CNN-DELM respectively.
with TensorFlow as the framework. MS Excel has been used Figure 9, and 10 shows the accuracy and loss curve for
for preparing required graphical representations of this study. the CNN. The training and testing accuracy for detection

VOLUME 11, 2023 4123

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

TABLE 3. Results of CNN-ELM with original images.

FIGURE 10. Loss curve of CNN with original images.

FIGURE 8. Confusion matrix for CNN-DELM with original images.

TABLE 4. Results of CNN-DELM with original images.

FIGURE 11. ROC curve of CNN with original images.

FIGURE 9. Accuracy curve of CNN with original images.

of malaria is 97.45%, and 96.59% respectively, whereas the

losses are 0.1126%, and 0.1472% respectively. All the scores
of CNN has been presented using Table 2.
The receiver operating characteristics (ROC) curve of the
CNN model is 98.98%, which showed in Figure 11. It serves
as a means of determining the efficacy of a learning model
for consistency [50]. After extracting the most prominent FIGURE 12. ROC curve of CNN-ELM with original images.
features using CNN, standardization has been performed on
these features. Then ELM has been applied for the detection.
First, CNN-ELM has been utilized and achieved an opti- CNN-DELM is 99.49% and 99.52%, respectively, as shown
mistic accuracy, precision, and recall of 97.68%, 98.09%, in Figures 12 and 13.
and 97.23%, respectively, which are shown in Table 3. Then
utilizing CNN-DELM, the model’s performance increased D. MODIFIED IMAGES
with accuracy and recall of 97.79%, and 97.67%, which are For modified images, the same methods have been utilized
demonstrated in Table 4. The ROC of the CNN-ELM and as performed in original images. It utilized 23543 images

4124 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 15. Confusion matrix of CNN with updated images.

FIGURE 13. ROC curve of CNN-DELM with original images.

FIGURE 16. Confusion matrix for CNN-ELM with updated images.

FIGURE 14. Graphical comparison for original images.

TABLE 5. Results of CNN with updates images.

FIGURE 17. Confusion matrix for CNN-DELM with updated images.

TABLE 7. Results of CNN-DELM with updated images.

TABLE 6. Results of CNN-ELM with updated images.

The accuracy and loss curves for the CNN are shown
in Figures 18, and 19. When it comes to the identification
of malaria, training and testing accuracy are 99.59% and
to train the models; 11759 have been from healthy people, 99.50%, with losses of 0.0065% and 0.0372%, respectively.
while 11784 have been from malaria-infected people. The All the scores of CNN with updated images have been pre-
suggested models’ performance and efficacy have been eval- sented in Table 5.
uated using 2616 data, including 1269 images of uninfected It is shown in Figure 20 that the CNN model has a ROC
cells and 1347 images of infected cells. The accuracy, pre- curve with a 99.99% confidence level. Following the extrac-
cision, recall, f1-score, and AUC of the proposed models tion of the most significant features using CNN, standard-
were calculated using a confusion matrix (CM). The CM for ization of these features has been conducted on the features
CNN, CNN-ELM and CNN-DELM have been presented in that have been derived. Then ELM and DELM were used
Figures 15, 16 and 17, respectively. to detect malaria, which was a breakthrough in the field.

VOLUME 11, 2023 4125

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

FIGURE 18. Accuracy curve of CNN with updated images.

FIGURE 21. ROC curve of CNN-ELM.

Then a CNN-DELM model has been introduced and achieved

an improved accuracy, and recall of 99.66%, which are repre-
sented in Table 7. According to Figures 21 and 22, the ROC
of the CNN-ELM and CNN-DELM is 99.49% and 99.52%,
respectively.

E. COMPARISON OF PERFORMANCE TO OTHER WORKS

The performance of the proposed DELM with shallow CNN
for malaria classification from processed RBC images has
been evaluated and compared with that of other current mod-
els. Here comparisons have been performed for original and
modified images. The current state-of-the-art models have
been summarised in Section II. Table 8 represents the compar-
ative analysis. Existing methods have used different train-test
FIGURE 19. Loss curve of CNN with updated images.
ratios. The proposed method has also been evaluated through
different train-test ratios to conduct the comparison. Based
on the ratios, the table is divided into four parts. First three
parts for the original dataset. The last part is for the modified
dataset.
Rajaraman et al. utilized the Kaggle RBC images for
the detection of malaria and achieved the highest accuracy
95.90% by using ResNet-50 [9]. Fatima et al. proposed a
computer-aided system for the prediction of malaria [51].
They have utilized four filters which were average, Gaus-
sian lowpass, median, bilateral filter for removing noises
and enhancing the quality of cell images. They achieved
high accuracy and f1-score of 91.80% and 91.53% using the
bilateral filter. Maqsood et al. trained their customized CNN
model using 17370 cell images and 8272 images used for
testing their model [16]. They achieved accuracy and f1-score
of 96.82%. The proposed CNN-DELM has been trained and
tested using the same fold of cell images from the original
FIGURE 20. ROC curve of CNN. dataset and achieved an optimistic accuracy and recall of
97.59%, and 97.58% respectively which are shown in table 8.
First and foremost, CNN-ELM has been employed, with Jain et al. used the Kaggle RBC images for trained their
optimistic accuracy, precision, and recall rates of 99.62%, various models [18]. They achieved the highest accuracy of
99.61%, and 99.62%, respectively, as indicated in Table 6. 97.00% while 9926 cell images using for testing their CNN

4126 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

TABLE 8. Comparative study of DELM model for malaria parasite

classification with other models.

FIGURE 22. ROC curve of CNN-DELM.

FIGURE 24. Graphical comparison of CNN-DELM with CNN [24].

for testing the proposed CNN-DELM and it has obtained

the highest accuracy and recall of 99.66%, and 99.55%
respectively.
Figure 24 shows the graphical comparison of the proposed
method with CNN [24]. From the comparative performance
FIGURE 23. Graphical comparison for updated images.
analysis represented in Table 8, the proposed model shows
its robustness. It outperforms the existing methods based
model. The proposed CNN-DELM model has been trained on different performance benchmarks. Table 9 shows differ-
and tested using the same dataset and achieved an accuracy ent state-of-the-art models and proposed models with brief
of 97.26%. structural descriptions. From Table 8 and Table 9, it can be
Khan et al. proposed three ML classifiers for the classi- said that the comparison is fair enough. The exact number
fication of malaria from cell images [22]. They tested their of test samples of a similar dataset (original and modified)
models using 5511 images from the Kaggle and achieved the has been used to compare. Data preprocessing (using the
highest accuracy recall and precision of 0.86, and 0.82 using proposed parasite inflator) and feature extraction techniques
a random forest classifier. The same testing dataset has been have been used, similar to the existing methods. Further, ELM
used for CNN-DELM and achieved recall, the precision of has been used as a classifier that acts as a strong classifier.
98.63% which is quite good from their work. The mixture of the feature extraction capability of CNN and
Fuhad et al. corrected the RBC images with help of a med- the excellent classification capability of DELM shows the
ical expert because some of the images are mislabeled [24]. way of optimistic outcomes. Hence, the unorthodox structure
they found 647 false parasitized and 750 false uninfected of the proposed model can claim novelty. The datasets that
images and they removed the falsely labeled images and have been used in this study, both original and updated, are
reduced the data from 27558 to 26161. They trained their balanced and large enough for the robustness of the model.
models CNN-SVM, CNN-KNN by using these modified There is color variation in the images of RBC from this
cell images. They have attained an accuracy 99.23% and dataset. As chemicals have been used to recognize malaria
recall of 99.54%. The same modified images have been used parasites in RBC, the color may change with the amount and

VOLUME 11, 2023 4127

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

TABLE 9. State-of-the-art models parameters in comparison with proposed model.

type or chemical. The model’s performance may vary with [5] M. P. Singh, K. B. Saha, S. K. Chand, and L. L. Sabin, ‘‘The economic
different colors from the existing images. cost of malaria at the household level in high and low transmission areas
of central India,’’ Acta Tropica, vol. 190, pp. 344–349, Feb. 2019.
In this study, images are good enough. The effectiveness of [6] M. O. F. Goni, F. M. S. Hasnain, M. A. I. Siddique, O. Jyoti, and
this framework with low quality is excluded from the scope M. H. Rahaman, ‘‘Breast cancer detection using deep neural network,’’ in
of this study. Proc. 23rd Int. Conf. Comput. Inf. Technol. (ICCIT), Dec. 2020, pp. 1–5.
[7] M. Goni, O. Faruq, A. Matin, T. Hasan, M. Sarker, and R. I. Islam,
‘‘Performance analysis of classifier for chronic kidney disease prediction
V. CONCLUSION using SVM, DNN and KNN,’’ in Proc. Int. Conf. Big Data, IoT, Mach.
in this paper a novel CNN-DELM has been presented for Learn. Singapore: Springer, 2022, pp. 3–13.
the automated diagnosis of malaria from tiny blood sam- [8] M. Nahiduzzaman, M. J. Nayeem, M. T. Ahmed, and M. S. U. Zaman,
‘‘Prediction of heart disease using multi-layer perceptron neural network
ples. The result shows that the detection accuracy improves and support vector machine,’’ in Proc. 4th Int. Conf. Electr. Inf. Commun.
when falsely labeled images are removed. Due to sev- Technol. (EICT), Dec. 2019, pp. 1–6.
eral morphological techniques, the malaria parasites are [9] S. Rajaraman, S. K. Antani, M. Poostchi, K. Silamut, M. A. Hossain,
highlighted. A lightweight CNN has been trained on prepro- R. J. Maude, S. Jaeger, and G. R. Thoma, ‘‘Pre-trained convolutional neu-
ral networks as feature extractors toward improved malaria parasite detec-
cessed cell images for extracting the most informative fea- tion in thin blood smear images,’’ PeerJ, vol. 6, Apr. 2018, Art. no. e4568.
tures. Eventually, DELM successfully differentiate between [10] A. Krizhevsky, I. Sutskever, and G. E. Hinton, ‘‘ImageNet classification
infected and uninfected samples from these features. The with deep convolutional neural networks,’’ in Proc. Adv. Neural Inf. Pro-
cess. Syst., vol. 25, 2012, pp. 1097–1105.
proposed parasitic inflator, CNN’s capability to extract fea-
[11] K. Simonyan and A. Zisserman, ‘‘Very deep convolutional networks for
tures, and DELM’s ability to generalize make the proposed large-scale image recognition,’’ 2014, arXiv:1409.1556.
approach more effective. The efficacy of the proposed frame- [12] K. He, X. Zhang, S. Ren, and J. Sun, ‘‘Deep residual learning for image
work is clear from its magnificent outcomes; whereas, the recognition,’’ in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR),
Jun. 2016, pp. 770–778.
CNN-DELM has obtained 97.79% accuracy and 97.88% [13] F. Chollet, ‘‘Xception: Deep learning with depthwise separable convolu-
recall for the original dataset. Also, it has acquired a promis- tions,’’ in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jul. 2017,
ing score of 99.66% accuracy and 99.55% recall for the pp. 1251–1258.
updated dataset. Since the proposed framework outperformed [14] M. Masud, H. Alhumyani, S. S. Alshamrani, O. Cheikhrouhou, S. Ibrahim,
G. Muhammad, M. S. Hossain, and M. Shorfuzzaman, ‘‘Leveraging
the state-of-art models, the result of this study is expected to deep learning techniques for malaria parasite detection using mobile
help medical professionals for detecting malaria and diagnose application,’’ Wireless Commun. Mobile Comput., vol. 2020, pp. 1–15,
the malaria-affected individuals more quickly and efficiently. Jul. 2020.
[15] L. Rosado, J. M. C. D. Costa, D. Elias, and J. S. Cardoso, ‘‘Automated
We have a plan to work on the multiple hidden layer ELM detection of malaria parasites on thick blood smears via mobile devices,’’
algorithm with CNN feature extractor and parasite inflator for Proc. Comput. Sci., vol. 90, pp. 138–144, Jan. 2016.
malaria detection. [16] A. Maqsood, M. S. Farid, M. H. Khan, and M. Grzegorzek, ‘‘Deep malaria
parasite detection in thin blood smear microscopic images,’’ Appl. Sci.,
vol. 11, no. 5, p. 2284, Mar. 2021.
REFERENCES [17] K. Sriporn, C.-F. Tsai, C.-E. Tsai, and P. Wang, ‘‘Analyzing malaria disease
using effective deep learning approach,’’ Diagnostics, vol. 10, no. 10,
[1] Malaria Microscopy Quality Assurance Manual—Version 2, World Health p. 744, Sep. 2020.
Org., Geneva, Switzerland, 2016.
[18] N. Jain, A. Chauhan, P. Tripathi, S. B. Moosa, P. Aggarwal, and B. Oznacar,
[2] W. R. J. Taylor, J. Hanson, G. D. H. Turner, N. J. White, and ‘‘Cell image analysis for malaria detection using deep convolutional net-
A. M. Dondorp, ‘‘Respiratory manifestations of malaria,’’ Chest, vol. 142, work,’’ Intell. Decis. Technol., vol. 14, no. 1, pp. 55–65, Mar. 2020.
no. 2, pp. 492–505, Aug. 2012. [19] Y. Dong, Z. Jiang, H. Shen, W. David Pan, L. A. Williams, V. V. B. Reddy,
[3] L. M. Barat, A. Mills, S. Basu, N. Palmer, K. Hanson, and E. Worrall, W. H. Benjamin, and A. W. Bryan, ‘‘Evaluations of deep convolutional
‘‘Do malaria control interventions reach the poor? A view through the neural networks for automatic identification of malaria infected cells,’’
equity lens,’’ Amer. J. Tropical Med. Hygiene, vol. 71, no. 2, pp. 174–178, in Proc. IEEE EMBS Int. Conf. Biomed. Health Informat. (BHI), 2017,
Aug. 2004. pp. 101–104.
[4] N. M. Gharakhanlou, M. S. Mesgari, and N. Hooshangi, ‘‘Developing [20] F. Yang, M. Poostchi, H. Yu, Z. Zhou, K. Silamut, J. Yu, R. J. Maude,
an agent-based model for simulating the dynamic spread of plasmodium S. Jaeger, and S. Antani, ‘‘Deep learning for smartphone-based malaria
vivax malaria: A case study of,’’ Ecol. Informat., vol. 54, Nov. 2019, parasite detection in thick blood smears,’’ IEEE J. Biomed. Health Infor-
Art. no. 101006. mat., vol. 24, no. 5, pp. 1427–1438, May 2020.

4128 VOLUME 11, 2023

M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

[21] D. Shah, K. Kawale, M. Shah, S. Randive, and R. Mapari, ‘‘Malaria feature extraction using extreme learning machine with CXR images,’’
parasite detection using deep learning: (Beneficial to humankind),’’ in IEEE Access, vol. 9, pp. 147512–147526, 2021.
Proc. 4th Int. Conf. Intell. Comput. Control Syst. (ICICCS), May 2020, [43] M. Nahiduzzaman, M. R. Islam, S. M. R. Islam, M. O. F. Goni,
pp. 984–988. M. S. Anower, and K.-S. Kwak, ‘‘Hybrid CNN-SVD based prominent fea-
[22] A. Khan, K. D. Gupta, D. Venugopal, and N. Kumar, ‘‘CIDMP: Completely ture extraction and selection for grading diabetic retinopathy using extreme
interpretable detection of malaria parasite in red blood cells using lower- learning machine algorithm,’’ IEEE Access, vol. 9, pp. 152261–152274,
dimensional feature space,’’ in Proc. Int. Joint Conf. Neural Netw. (IJCNN), 2021.
Jul. 2020, pp. 1–8. [44] S. Ding, N. Zhang, X. Xu, L. Guo, and J. Zhang, ‘‘Deep extreme learning
[23] A. Olugboja and Z. Wang, ‘‘Malaria parasite detection using different machine and its application in EEG classification,’’ Math. Problems Eng.,
machine learning classifier,’’ in Proc. Int. Conf. Mach. Learn. Cybern. vol. 2015, pp. 1–11, Nov. 2015.
(ICMLC), Jul. 2017, pp. 246–250. [45] D. Xiao, B. Li, and Y. Mao, ‘‘A multiple hidden layers extreme learning
[24] K. M. F. Fuhad, J. F. Tuba, M. R. A. Sarker, S. Momen, N. Mohammed, machine method and its application,’’ Math. Problems Eng., vol. 2017,
and T. Rahman, ‘‘Deep learning based automatic malaria parasite detection pp. 1–10, Jan. 2017.
from blood smear and its smartphone based application,’’ Diagnostics, [46] D. M. W. Powers, ‘‘What the F-measure doesn’t measure: Features, flaws,
vol. 10, no. 5, p. 329, May 2020. fallacies and fixes,’’ 2015, arXiv:1503.06410.
[25] D. Anggraini, A. S. Nugroho, C. Pratama, I. E. Rozi, A. A. Iskandar, and [47] L. V. R. Asuncion, J. X. P. D. Mesa, P. K. H. Juan, N. T. Sayson, and
R. N. Hartono, ‘‘Automated status identification of microscopic images A. R. D. Cruz, ‘‘Thigh motion-based gait analysis for human identifica-
obtained from malaria thin blood smears,’’ in Proc. Int. Conf. Electr. Eng. tion using inertial measurement units (IMUs),’’ in Proc. IEEE 10th Int.
Informat., Jul. 2011, pp. 1–6. Conf. Humanoid, Nanotechnol., Inf. Technology,Communication Control,
[26] G. P. Gopakumar, M. Swetha, G. S. Siva, and G. R. K. S. Subrahmanyam, Environ. Manage. (HNICEM), Nov. 2018, pp. 1–6.
‘‘Convolutional neural network-based malaria diagnosis from focus stack [48] D. M. W. Powers, ‘‘Evaluation: From precision, recall and
of blood smear images acquired using custom-built slide scanner,’’ J. Bio- F-measure to ROC, informedness, markedness and correlation,’’ 2020,
photon., vol. 11, no. 3, Mar. 2018, Art. no. e201700003. arXiv:2010.16061.
[27] Z. Liang, A. Powell, I. Ersoy, M. Poostchi, K. Silamut, K. Palaniappan, [49] Y. Sasaki and R. Fellow, ‘‘The truth of the F-measure,’’ MIB-School
P. Guo, M. A. Hossain, A. Sameer, R. J. Maude, J. X. Huang, S. Jaeger, and Comput. Sci., Univ. Manchester, Manchester, U.K., Tech. Rep. 5, 2007,
G. Thoma, ‘‘CNN-based image analysis for malaria diagnosis,’’ in Proc. p. 25.
IEEE Int. Conf. Bioinf. Biomed. (BIBM), Dec. 2016, pp. 493–496. [50] D. J. Peres, C. Iuppa, L. Cavallaro, A. Cancelliere, and E. Foti, ‘‘Significant
wave height record extension by neural networks and reanalysis wind
[28] R. Tomari, W. N. W. Zakaria, M. M. A. Jamil, F. M. Nor, and N. F. N. Fuad,
data,’’ Ocean Model., vol. 94, pp. 128–140, Oct. 2015.
‘‘Computer aided system for red blood cell classification in blood smear
[51] T. Fatima and M. S. Farid, ‘‘Automatic detection of plasmodium parasites
image,’’ Proc. Comput. Sci., vol. 42, pp. 206–213, Jan. 2014.
from microscopic blood images,’’ J. Parasitic Diseases, vol. 44, no. 1,
[29] G. Díaz, F. A. González, and E. Romero, ‘‘A semi-automatic method
pp. 69–78, Mar. 2020.
for quantification and classification of erythrocytes infected with malaria
[52] M. Hasan, S. Islam, A. Dey, A. Das, and S. C. Tista, ‘‘Detection of malaria
parasites in microscopic images,’’ J. Biomed. Informat., vol. 42, no. 2,
disease using image processing and machine learning,’’ in Proc. Int. Conf.
pp. 296–307, Apr. 2009.
Intell. Comput. Optim. Cham, Switzerland: Springer, 2021, pp. 457–466.
[30] P. A. Pattanaik, M. Mittal, and M. Z. Khan, ‘‘Unsupervised deep learning
[53] E. T. Efaz, F. Alam, and M. Kamal, ‘‘Deep CNN-supported ensemble
CAD scheme for the detection of malaria in blood smear microscopic
CADx architecture to diagnose malaria by medical image,’’ in Proc.
images,’’ IEEE Access, vol. 8, pp. 94936–94946, 2020.
Int. Conf. Trends Comput. Cognit. Eng. Singapore: Springer, 2021,
[31] D. Bibin, M. S. Nair, and P. Punitha, ‘‘Malaria parasite detection from pp. 231–243.
peripheral blood smear images using deep belief networks,’’ IEEE Access, [54] D. K. Das, M. Ghosh, M. Pal, A. K. Maiti, and C. Chakraborty, ‘‘Machine
vol. 5, pp. 9099–9108, 2017. learning approach for automated screening of malaria parasite using light
[32] Arunava. (2018). Malaria Cell Images Dataset. Accessed: Aug. 10, 2021. microscopic images,’’ Micron, vol. 45, pp. 97–106, Feb. 2013.
[Online]. Available: https://www.kaggle.com/iarunava/cell-images-for- [55] M. Rottmann et al., ‘‘Spiroindolones, a potent compound class for the treat-
detecting-malaria ment of malaria,’’ Science, vol. 329, no. 5996, pp. 1175–1180, Sep. 2010.
[33] (2019). Corrected Malaria Data—Google Drive. Accessed: [56] F. J. P. Montalbo and A. S. Alon, ‘‘Empirical analysis of a fine-tuned deep
Aug. 10, 2021. [Online]. Available: https://drive.google.com/drive/folders/ convolutional model in classifying and detecting malaria parasites from
10TXXa6B_D4AKuBV085tX7UudH1hINBRJ?usp=sharing blood smears,’’ KSII Trans. Internet Inf. Syst., vol. 15, no. 1, pp. 147–165,
[34] S. Sarangi, M. Sahidullah, and G. Saha, ‘‘Optimization of data-driven 2021.
filterbank for automatic speaker verification,’’ Digit. Signal Process.,
vol. 104, Sep. 2020, Art. no. 102795.
[35] S. Ioffe and C. Szegedy, ‘‘Batch normalization: Accelerating deep network
training by reducing internal covariate shift,’’ in Proc. Int. Conf. Mach.
Learn., 2015, pp. 448–456.
[36] D. Ciresan, U. Meier, and J. Schmidhuber, ‘‘Multi-column deep neural
networks for image classification,’’ in Proc. IEEE Conf. Comput. Vis.
Pattern Recognit., Jun. 2012, pp. 3642–3649.
[37] R. Jain, P. Nagrath, G. Kataria, V. S. Kaushik, and D. J. Hemanth, ‘‘Pneu-
monia detection in chest X-ray images using convolutional neural networks
and transfer learning,’’ Measurement, vol. 165, Dec. 2020, Art. no. 108046.
[38] S. Mittal, ‘‘A survey of FPGA-based accelerators for convolutional neu-
ral networks,’’ Neural Comput. Appl., vol. 32, no. 4, pp. 1109–1139,
Feb. 2020.
[39] G. Kovács, L. Tóth, D. Van Compernolle, and S. Ganapathy, ‘‘Increasing
MD. OMAER FARUQ GONI received the
the robustness of CNN acoustic models using autoregressive moving aver-
age spectrogram features and channel dropout,’’ Pattern Recognit. Lett., B.Sc. degree in electronics and communication
vol. 100, pp. 44–50, Dec. 2017. engineering (ECE) from the Rajshahi Univer-
[40] D. P. Kingma and J. Ba, ‘‘Adam: A method for stochastic optimization,’’ sity of Engineering and Technology (RUET),
2014, arXiv:1412.6980. Bangladesh, Kazla, Rajshahi, in 2021, where he is
[41] G.-B. Huang, Q.-Y. Zhu, and C.-K. Siew, ‘‘Extreme learning machine: The- currently pursuing the M.Sc. degree in computer
ory and applications,’’ Neurocomputing, vol. 70, nos. 1–3, pp. 489–501, science and engineering (CSE). He is also a Lec-
2006. turer with the Department of Electrical and Com-
[42] M. Nahiduzzaman, M. O. F. Goni, M. S. Anower, M. R. Islam, M. Ahsan, puter Engineering (ECE), RUET. His research
J. Haider, S. Gurusamy, R. Hassan, and M. R. Islam, ‘‘A novel method for interests include deep learning, power system
multivariant pneumonia classification based on hybrid CNN-PCA based automation, and the Internet of Things.

VOLUME 11, 2023 4129

 M. Omaer Faruq Goni et al.: Diagnosis of Malaria Using Double Hidden Layer ELM Algorithm

MD. NAZRUL ISLAM MONDAL (Senior Mem- MD. NAHIDUZZAMAN received the B.Sc.
ber, IEEE) received the B.E. degree in electrical degree in computer science and engineering from
and electronic engineering from the Department the Rajshahi University of Engineering and Tech-
of Electrical and Electronic Engineering, Royal nology (RUET), Kazla, Rajshahi, Bangladesh,
University of Engineering and Technology, Dhaka, in 2018. He is currently a Lecturer with the
Bangladesh, in 2000, the M.E. degree in informa- Department of Electrical and Computer Engineer-
tion and communication technologies (ICT) from ing, RUET. Apart from this, he also works as
the School of Engineering and Technology, Asian a Research Assistant with the Qatar University
Institute of Technology, Bangkok, in 2008, and Machine Learning Research Group. He has several
the Ph.D. degree in information engineering from peer-reviewed journal publications. His research
Hiroshima University, Japan, in 2012. interests include machine learning and its applications in disease detection,
In 2001, he worked as a Lecturer at the Department of Computer power sector, and agriculture. He is a Reviewer of BioMedical Engineering
Science and Technology, Rajshahi University of Engineering and Technol- OnLine.
ogy (RUET). He was promoted as an Assistant Professor at the Depart-
ment of Computer Science and Technology, in 2004. From March 2012 to
MD. ROBIUL ISLAM received the B.Sc. degree
September 2012, he was a Visiting Research Scholar at the Department
in computer science and engineering from the
of Information Engineering, Hiroshima University. He was also promoted
Rajshahi University of Engineering and Tech-
as an Associate Professor with the Department of Computer Science and
nology (RUET), Kazla, Rajshahi, Bangladesh,
Technology, in 2013. From May 2013 to April 2014, he was a Specially
in 2017, where he is currently pursuing the M.Sc.
Appointed Visiting Lecturer at the Department of Information Engineering,
degree. He is also an Assistant Professor with the
Hiroshima University, where he was also a Visiting Researcher. In June 2015,
Department of Electrical and Computer Engineer-
he was hired as an Associate Professor with the Department of Information
ing, RUET. As a data scientist, he is fascinated
Engineering. As a Cisco Instructor, he has been working with students,
by topics like segmentation, machine learning, and
since 2006. He is currently a Professor with the Department of Computer
pattern recognition.
Science and Engineering, RUET. He has a long list of publications in journals
and conference proceedings to his credit. There are numerous journals and
conferences where his work can be found in the form of organizing chairman- MD. SHAMIM ANOWER (Member, IEEE)
ships, membership on PCs, reviews, and sub-reviews. These include journals received the Ph.D. degree (Hons.) in electrical
such as IEICE, the Journal of Foundational Computer Science, the Journal engineering from the University of New South
of Communication/Computer/Information Technology, and the International Wales, in 2012. He is currently a Professor with
Journal of Networking and Computing. He is a member of the Institution the Department of Electrical and Electronic Engi-
of Engineers, Bangladesh (F/10973). He has organized several well-known neering, Rajshahi University of Engineering and
conferences, workshops, and programming competitions at the international Technology. He has more than 150 peer-reviewed
and national levels. publications to his credit, including journal
(18 Q1 Scimago ranked) and conference articles
with H-indices of 20, i10-indices of 33, and
1205 citations. Energy and mining technology, quantum information,
advanced digital, data science, and ICT, and cyber security are the three
focus areas covered by these reports. He is also an IEB Fellow.

KYUNG-SUP KWAK (Life Senior Member,

IEEE) received the Ph.D. degree from the
University of California. He was with Hughes
Network Systems and the IBM Network Analy-
sis Center, USA. He was with Inha University,
S. M. RIAZUL ISLAM (Member, IEEE) received South Korea, as a Professor. He was also the Dean
the Ph.D. degree in information engineering of the Graduate School of Information Technology
from Inha University, South Korea. Before mov- and Telecommunications and the Director of the
ing to the U.K., he was an Assistant Pro- UWB Wireless Communications Research Center.
fessor at the Department of Computer Science In 2008, he was an Inha Fellow Professor (IFP).
and Engineering, Sejong University, South Korea, He is currently an Inha Hanlim Fellow Professor and a Professor with the
from 2017 to 2022. He was a Postdoctoral Fellow School of Information and Communication Engineering, Inha University.
at Inha University, a Senior Engineer at Samsung His research interests include UWB radio systems, wireless body area
Research and Development Institute, and an Assis- networks and u-health networks, and nano and molecular communications.
tant Professor of electrical and electronics engi- In 2006, he was the President of the Korean Institute of Communication
neering at the University of Dhaka. He is currently a Senior Lecturer in Sciences (KICS) and the Korea Institute of Intelligent Transport Systems
computer science with the University of Huddersfield, U.K. He has authored (KITS), in 2009. He received the official commendations for achievements
more than 100 publications in peer-reviewed international journals. His of UWB radio technology research and development from the Korean
research interests include applied AI, machine learning, data science, and President, in 2009.
the IoT.

4130 VOLUME 11, 2023