An Effective and Novel Approach For Brain Tumor CL

Hindawi
Journal of Sensors
Volume 2023, Article ID 1224619, 19 pages
https://doi.org/10.1155/2023/1224619
Research Article
An Effective and Novel Approach for Brain Tumor Classification
Using AlexNet CNN Feature Extractor and Multiple Eminent
Machine Learning Classifiers in MRIs
Alok Sarkar ,1 Md. Maniruzzaman ,1 Mohammad Ashik Alahe ,1

and Mohiuddin Ahmad 2
1
Electronics and Communication Engineering Discipline, Khulna University, Khulna, 9208, Bangladesh
2
Department of Electrical and Electronic Engineering, Khulna University of Engineering & Technology, Khulna, 9203, Bangladesh
Correspondence should be addressed to Alok Sarkar; [email protected]
Received 27 October 2022; Revised 28 January 2023; Accepted 27 February 2023; Published 8 March 2023
Academic Editor: Akhilesh Pathak
Copyright © 2023 Alok Sarkar et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A brain tumor is an uncontrolled malignant cell growth in the brain, which is denoted as one of the deadliest types of cancer in
people of all ages. Early detection of brain tumors is needed to get proper and accurate treatment. Recently, deep learning
technology has attained much attraction to the physicians for the diagnosis and treatment of brain tumors. This research
presents a novel and effective brain tumor classification approach from MRIs utilizing AlexNet CNN for separating the dataset
into training and test data along with extracting the features. The extracted features are then fed to BayesNet, sequential
minimal optimization (SMO), Naïve Bayes (NB), and random forest (RF) classifiers for classifying brain tumors as no-tumor,
glioma, meningioma, and pituitary tumors. To evaluate our model’s performance, we have utilized a publicly available Kaggle
dataset. This paper demonstrates ROC, PRC, and cost curves for realizing classification performance of the models; also,
performance evaluating parameters, such as accuracy, sensitivity, specificity, false positive rate, false negative rate, precision, f-
measure, kappa statistics, MCC, ROC area, and PRC area, have been calculated for four testing options: the test data itself,
cross-validation fold (CVF) 4, CVF 10, and percentage split (PS) 34% of the test data. We have achieved 88.75%, 98.15%,
86.25% and 100% of accuracy using the AlexNet CNN+BayesNet, AlexNet CNN+SMO, AlexNet CNN+NB, and AlexNet CNN
+RF models, respectively, for the test data itself. The results imply that our approach is outstanding and very effective.
1. Introduction structure of the brain to diagnose tumors. The most popular

imaging techniques are ultrasound imaging, X-ray imaging,
The existence of any kind of abnormalities in the brain positron emission tomography, computed tomography,
may lead the human nervous system as well as health in and magnetic resonance imaging (MRI). Among these, the
great danger [1]. Today, brain tumor is denoted as the MRI provides precise information on a given medical image,
most severe ones, which results from the development of gives definite information on the shape of the cerebrum
uncontrolled destructive cells in the human brain. Accord- structure, and the location of inconsistencies in brain tissues.
ing to the World Health Organization and the American However, it is often chaotic, time-consuming, and error-
Brain Tumor Association reports [2], it can be categorized prone if medical specialists manually detect brain tumors
as glioma, meningioma and pituitary tumors based on cell by examining MRIs [3]. To diminish these problems, the
structure and locations. computer-aided detection and diagnostic approach has been
It is necessary to completely identify the types of brain developed for the last few decades [4–11]. In recent years,
tumors for proper diagnosis and treatment. The radiologist various efforts have been taken to elevate a robust and
utilizes various clinical imaging techniques to visualize the precise technique to classify brain tumors automatically by
2 Journal of Sensors
using machine learning in MRIs [12–14]. Nevertheless, second-order statistics and a least square support vector
because of high inter- and intrashape, texture, and variation machine (SVM) classifier. Sarkar et al. [38] proposed a
of contrast, this still remains a challenging task. Conven- computer-aided approach for detecting and classifying brain
tional machine learning techniques rely on artisanal features, tumors from MRIs utilizing genetic algorithm as a feature
which limit the robustness of the outcomes, while deep extractor and SVM as classifier, where they obtained 98.3%
learning techniques automatically extract important func- classification accuracy. John [25] used techniques based on
tionality that provides significantly better performance. the gray level cooccurrence matrix (GLCM) and discrete
But still, the results obtained using deep learning tech- wavelet transform (DWT) to detect and classify brain
nologies are not sufficient, and they do not always show tumors. Ullah et al. [39] applied DWT for feature extraction
good classification accuracy, uses of only AlexNet, for exam- and a feed-forward artificial neural network for brain MRI
ple [15]. The research is challenging, especially for selecting classification. Kharrat et al. [40] classified brain tumors as nor-
the region of interest (ROI), extracting deep features, using mal and abnormal classes through genetic algorithms and
clamorous images, etc. SVM classifiers. Papageorgiou et al. [33] obtained 90.26%
To minimize the aforementioned problems, in this and 93.22% accuracy for low- and high-grade glioma tumors,
study, we have proposed a novel and effective technique respectively, through fuzzy cognitive maps. Díaz-Pernas et al.
for brain tumor classification automatically employing the [26] proposed an architecture for automatic detection and seg-
AlexNet CNN feature extractor and a variety of prominent mentation of brain tumors and acquired 97.3% accuracy.
machine learning classifiers, such as BayesNet, sequential Shree and Kumar [41] used a GLCM feature extractor and
minimal optimization (SMO), Naïve Bayes (NB), and ran- probabilistic neural network classifier for classifying MRI
dom forest (RF) in MRIs. datasets and achieved 95% accuracy. Arunachalam and Savar-
We know that AlexNet CNN is a good tool to determine imuthu [42] offered an architecture that covered enhancement
ROI for extracting deep features and can easily deal with of image, transformation of image, extraction of features, and
clamorous and raw images, which are very essential for the finally, image classification. They employed shift-invariant
current research. Moreover, deploying machine learning shear-let transforms for enhancing MRIs and used numerous
classification algorithms with AlexNet CNN certainly feature extractors, such as Gabor, GLCM, and DWT to extract
enhance the performance of the proposed model. the features. Hossain et al. [30] utilized the traditional CNN
Our contributions to this work are as follows: approach to classify brain tumor images. They utilized fuzzy
c-means clustering for segmenting the tumors and achieved
(i) We have utilized AlexNet CNN for separating an accuracy of 97.87%.
human brain MRI datasets into training and test Several deep learning approaches have been extensively
data and also for extracting the features from those utilized by researchers to detect and classify brain tumors
separated datasets in the last decades. Sajid et al. [43] proposed a fully auto-
mated hybrid approach for the detection and segmentation
(ii) Those extracted features are then used for classifying
of brain tumors in MRIs using deep learning. They utilized
the brain tumors as no-tumor, glioma, meningioma,
the BRATS2013 dataset and obtained 86%, 86%, and 91%
and pituitary tumors by using eminent BayesNet,
dice scores, sensitivity, and specificity, respectively. Saxena
SMO, NB, and RF classifiers
et al. [44] utilized ResNet-50, Inception V3, VGG-16, and
This paper is organized as follows. Section 1 discusses transfer learning techniques to classify the images; the
briefly the problems, some of the existing solutions, research ResNet-50 outperformed with 95% accuracy. In [27], Çinar
gaps, objectives, contributions, and organizations of the paper. and Yildirim utilized numerous convolutional neural
Section 2 presents related works on brain tumor detection and network (CNN) models, such as Inception V3, GoogLeNet,
classification. It also discusses the research gaps at the end of ResNet-50, AlexNet, and DenseNet-201, for classifying the
this section. Section 3 describes overall working methodology, MRI datasets; whether they acquired respectable accuracies
proposed architecture, and performance measuring metrics. in every case. Khwaldeh et al. [45] customized the AlexNet
Section 4 describes the results and performance of our model architecture and employed it to classify brain MRIs, and
along with a comparison with other contemporary findings. obtained 91% accuracy. Preethi and Aishwarya [46]
Finally, Section 5 contains the conclusion of the entire study, presented an architecture that combined wavelet-based
limitations, and future works. GLCM to develop matrices of features. They utilized a deep
neural network classifier to detect and categorize brain MRIs
and obtained 92% accuracy. In [28], Hemanth et al. pro-
2. Related Works posed a modified deep convolutional neural network
(DCNN) to classify brain tumors, and they obtained excel-
Many researchers have published their works to represent lent model performance in terms of accuracy. Khan et al.
different methodologies or approaches for detection and [47] introduced a computerized multimodel categorization
classification purposes over the past few years [15–36]. technique to detect and classify brain tumor images. They
Various methodologies have been tested using different clin- utilized two presupervised CNN models, known as VGG16
ical databases, including magnetic resonance imaging (MRI) and VGG19, to extract features. They deployed the
of brain tumors. Selvaraj et al. [37] developed a binary BRATS2018 dataset to validate their model and obtained
classifier for classifying brain MRIs a utilizing first- and about 97.80% accuracy.
Journal of Sensors 3
Input MR images
Training image set Test image set
Preprocessing Preprocessing
Feature extraction Feature extraction
Extracted features Extracted features
Classifier
Prediction model (BayesNet, SMO,
NB, RF)
Classified result
Glioma Meningioma No-tumor Pituitary
Figure 1: Block diagram of the proposed method for identification and classification of brain tumors.
accuracy was obtained. Updated versions of CNN architec-

Table 1: Statistical description of data.
tures, namely, AlexNet and VGG16, were introduced by
Image type Total images Training images Testing images Ali et al. [15] with 96% and 98% of accuracy, respectively.
Glioma 900 630 270
Deepak and Ameer [32] proposed a fully computerized
technique for classifying brain tumor images into three
Meningioma 900 630 270
categories: glioma, meningioma, and pituitary tumors.
No-tumor 900 630 270 Their designed architecture was an updated version of
Pituitary 900 630 270 the GoogLeNet model and obtained 98% accuracy. In
Total 3600 2520 1080 another study, Sultan et al. [34] presented a DCNN model
for accurately classifying MRI data into glioma, meningi-
oma, and pituitary tumors, where they achieved 96% accu-
racy. Abiwinanda et al. [35] designed numerous CNN
In [48], Siar and Teshnehlab utilized a central cluster- architectures for classifying brain tumors into three classes:
ing algorithm to extract features, then fed those features glioma, meningioma, and pituitary tumors. One of the
directly to the CNN model and achieved 96% accuracy. architectures showed the best result with 98.51% of accu-
Kuang et al. [49] introduced a 3D CNN model to classify racy. In [36], Seetha and Raja classified the brain tumor
the BRATS2018 brain tumor image dataset. Preprocessing images using the CNN. They extracted the features using
was performed using the grey-level normalization tech- ImageNet and obtained 97.50% accuracy.
nique and contrast adjustment. Afterward, those images Every method has some advantages and disadvantages.
were transferred to the 3D CNN model for classification. Some research gaps are identified in the above-mentioned
They achieved a dice score of 92%. In another study papers and summarized briefly as follows.
[29], Zhou et al. considered 3D images as 2D cross-
sections to utilize those images as the input dataset. They (i) Many researchers have used different techniques in
employed different models, such as DenseNet-RNN, Den- particular to determine the region of interest
seNet-LSTM, and DenseNet-DenseNet, and obtained accu- (ROI) for extracting features
racies of 87%, 91% and 92%, respectively. In [31],
Krishnammal and Raja utilized a presupervised AlexNet (ii) Many researchers have used statistical features and
model to classify the MRI dataset. Features were extracted texture features, but these have not given so much
through curvelet transform and GLCM matrix, and 100% good accuracy with respect to deep features
Fully connected layer
Convolution + Convolution +
Output
ReLU Pooling ReLU Pooling
Glioma
Meningioma
No-tumor
•••••
Pituitary
Figure 2: Basic CNN architecture.
1st CON 11×11, 55×55×96

227×227×3 Stride = 4 55×55×96 2nd CON 5×5,
Max pool 3×3 3rd CON 3×3,
Kemels = 96 Padding = 2, 27×27×256 Max pool 3×3, Padding = 1,
Stride = 2 Kemels = 256 Stride = 2 13×13×256 Kemels = 384 13×13×384
55×55 55×55
227×227 {(227-11)/4}+1=55 27×27
{(55-3)/2}+1=27 ⁎
{(27+2 2-5)/1}+1=27 {(27-3)/2}+1=13 13×13 ⁎
13×13
{(13+2 1-3)/1}+1=13
{(13+2⁎1-3)/1}+1=13
4th CON 3×3,
Kemels = 384
Padding = 1,
1000 softmax Fully connected 8 13×13 13×13
Fully connected 7
6×6
Fully connected 6 {(13-3)/2}+1=6 ⁎
{(13+2 1-3)/1}+1=13
6×6×256 Max pool 3×3, 5th CON 3×3,

Stride = 2 13×13×256 Padding = 1, 13×13×384
Kemels = 256
4096 4096
Figure 3: The proposed architecture of AlexNet.
(iii) Most of the studies have used good quality MRI rated in the upcoming subsections. We have used AlexNet
images. Raw images or noisy images have been CNN in the MATLAB platform [50] for categorizing the
found well dealt by deep learning methods dataset as training and test as well as extracting the features
from it. Then, the extracted features are employed in the
(iv) Manual classification of brain tumors from MRIs is WEKA platform [51] to classify the brain tumors of MRIs
time-consuming, nonreproducible, and impractical as no-tumor, glioma, meningioma, and pituitary tumors uti-
if the data size is larger lizing prominent BayesNet, SMO, NB, and RF classifiers and
Thus, an excellent model is required for feature extraction establishing a prediction model. The performance of the
and classification of brain tumors accurately from MRIs. model is assessed by the classified results.
3. Materials and Methods 3.2. Data Description. The selection of an appropriate data-
set is the first and foremost concern for identifying and clas-
3.1. Research Implementation Block Diagram. Figure 1 shows sifying medical images. Therefore, we collected the dataset
the overall working architecture of our proposed model to from a trustworthy website, the Kaggle database [52, 53], a
accurately identify and classify brain tumors from MRIs. well-known web-based data source for brain MRIs. From
Each key component of this architecture, such as the dataset [52], we have utilized the glioma, meningioma, and pituitary
including training and test data, preprocessing, feature tumor data, and from [53], we have utilized the no tumor
extraction, prediction model, classified results, and perfor- data. In total, 3600 images have been used in this study,
mance measurement of the diagram, are sequentially nar- where each class contains the same number of images (900
Table 2: Confusion matrix for (a) test data itself, (b) CVF 4, (c) CVF 10, and (d) PS 34% of test data using BayesNet, SMO, NB, and RF classifiers.
(a)
BayesNet classifier SMO classifier

Glioma Meningioma No tumor Pituitary Glioma Meningioma No tumor Pituitary
Journal of Sensors
Glioma 203 36 6 25 Glioma 252 17 1 0

Meningioma 30 175 27 38 Meningioma 6 256 3 5
No-tumor 5 22 243 0 No-tumor 0 1 269 0
Pituitary 14 37 3 216 Pituitary 3 4 0 263
NB classifier RF classifier
Glioma 197 32 6 35 Glioma 270 0 0 0
No-tumor 5 29 235 1 No-tumor 0 0 270 0
(b)

Glioma 181 51 9 29 Glioma 224 34 3 9
No-tumor 6 31 233 0 No-tumor 5 11 251 3
Glioma 191 35 8 36 Glioma 207 45 4 14
No-tumor 5 29 235 1 No-tumor 3 10 256 1
(c)

Glioma 183 50 7 30 Glioma 211 44 5 10
No-tumor 8 30 232 0 No-tumor 4 9 254 3
5
6
Table 2: Continued.

Glioma 189 38 6 37 Glioma 214 36 5 15
No-tumor 5 29 235 1 No-tumor 4 9 256 1
(d)

Glioma 116 37 5 26 Glioma 132 38 4 10
No-tumor 6 17 152 0 No-tumor 3 9 161 2
Glioma 120 36 5 23 Glioma 134 35 5 10
No-tumor 6 15 154 0 No-tumor 2 9 163 1
Journal of Sensors
1
1
0.8
0.8
True positive rate
True positive rate

0.6
0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
False positive rate False positive rate
BayesNet NB BayesNet NB
SMO RF SMO RF
(a) (b)
1 1
0.8 0.8
True positive rate
True positive rate
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1

SMO RF SMO RF
(c) (d)
Figure 4: ROC curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for test data itself using BayesNet, SMO, NB,
and RF classifiers.
per class). The elaborate allocation of the data collected and clamor. Therefore, it is important to evacuate the conceiv-
deployed in this study is shown in Table 1. We split the data- able clamors through preprocessing. In addition, enhance-
set (3600 MRIs) into training and test datasets: 2520 (70% of ment of the visual nature of images offers tremendous
the total dataset) images are utilized as training images, and favor to the specialists. In this research, MRIs are prepro-
the rest 1080 (30% of the total dataset) images as test images, cessed with an anisotropic filter. The size of the images in
using AlexNet CNN. Training data are deployed for model the dataset of MRIs is not the same. These different sizes
prediction; on the other hand, test data are applied for testi- of images illustrate the input layer of the CNN network;
fying the model’s performance. therefore, these images need to be preprocessed, standard-
ized, and resized to 227 × 227 pixels.
3.3. Preprocessing. Preprocessing is a technique that aug-
ments the quality of images and makes them able for further 3.4. Feature Extraction. One presupervised CNN model,
steps. In a certain cases, the separation between normal and AlexNet [15, 31], is utilized in this investigation to include
abnormal tissue is complicated because of the high-clamor extraction, because it is a fantastically incredible model fit
level. As a result, specialists may normally commit errors for accomplishing high-level accuracy on testing datasets.
in diagnosis. Then again, minor contrasts between normal AlexNet is the most important known deep CNN structure
and abnormal tissues can likewise be concealed by the which comprises twenty-five layers, eight of which add to
1 1
0.9 0.9
0.8 0.8
0.7 0.7
True positive rate
True positive rate

0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1
0.9
0.8 0.8
0.7
True positive rate
True positive rate
0.6 0.6
0.5
0.4 0.4
0.3
0.2 0.2
0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 5: ROC curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for CVF 4 of the test data using BayesNet,
SMO, NB, and RF classifiers.
learning by modifying loads. Five of these are known as con- future results or conducts are predicted dependent on the
volution layers while staying three are known as fully con- former and present data. It is a measurable examination pro-
nected (FC) layers. In the AlexNet design, max-pooling cedure that empowers the assessment and count of the like-
layers are connected to convolution layers in series with each lihood of specific outcomes identified with programming or
other. The first convolution layer and then max-pooling frameworks. The prediction model works by gathering
layers utilize fluctuating kernel sizes [54]. The layer of information, making a factual model, and applying probabi-
max-pooling ensues the layer of convolution [54, 55]. FC6 listic methods to anticipate the possible results.
is the first FC layer, and FC7 is the second FC layer in acti-
vation are utilized to remove the features of vectors. There 3.6. Proposed Network Architecture
are altogether 4096 features of FC6 and FC7 vectors in the
AlexNet CNN design [56]. 3.6.1. Convolutional Neural Network. The convolutional
neural network (CNN) is a sort of multilayered feed-
3.5. Prediction Model. Consequently, we have developed pre- forward neural network and one of the significant ideas of
dictive models based on features extracted from the trained the deep learning method. Its basic structure consists of sev-
data. A prediction model is a procedure through which eral convolution layers, a rectified linear unit (ReLU) layer,
1 1
0.8 0.8
True positive rate
True positive rate

0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1
0.8 0.8
True positive rate
True positive rate
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 6: ROC curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for the CVF 10 of test data using BayesNet,
SMO, NB and RF classifiers.
pooling layers, and FC layers, as shown in Figure 2. The (11 × 11, 5 × 5, 3 × 3, 3 × 3, 3 × 3), three max-pooling layers
basic function of the convolution layer is to extract the (3 × 3), and three FC layers. The first two max-pooling
local features of the existing layers. The ReLU layer per- layers are deployed after the first two convolutional layers
forms element-by-element activation. The pooling layers sequentially. The third, fourth, and fifth convolutional
are introduced for downsampling. Max-pooling is com- layers are connected directly. The third max-pooling layer
monly utilized to reduce the accuracy of the feature maps is inserted after the fifth convolutional layer, and its out-
by incorporating linguistically similar features. To over- put is fed into a series of three FC layers. The success of
come the fitting issues, some neurons have been removed AlexNet is usually credited to ReLU, stochastic gradient
from the CNN architecture, known as dropouts. At the descent (SGD), dropout, etc. ReLU is introduced for accel-
end, the FC layer yields the class score value ranging from erating the speed of training processes. The value of the
0 to 1 which is deployed to attain the classification deci- convolutional kernel is extracted by optimizing the total
sions. Essentially, SoftMax layers are introduced by default cost function and applying the SGD algorithm. To elimi-
in this motive. nate the overfitting problem, a dropout layer is applied
in the first two FC layers. The third FC layer, also known
3.6.2. AlexNet. Figure 3 shows the architecture of an AlexNet as the softmax layer, is used to classify various objects in a
which is mainly comprised of five convolutional layers CNN [57].
1 1
0.8 0.8
True positive rate
True positive rate

0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1
0.9
0.8 0.8
0.7
True positive rate
True positive rate
0.6 0.6
0.5
0.4 0.4
0.3
0.2 0.2
0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 7: ROC curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for PS 34% of the test data using BayesNet,
3.7. Classifiers be materialized by training the SVMs. At this issue, the

least plausible optimization problem includes the Lagrange
3.7.1. BayesNet Classifier. The BayesNet classifier is a kind of multipliers which have two numbers, because the Lagrange
directed acyclic graph which enciphers a joint probability multipliers must comply with a direct equity limitation.
distribution over a set of random variables [58, 59]. It is a First, it discovers two multipliers, and then it attempts to
twosome of B = ðG, θÞ, where G represents the graph whose enhance them. It will be repeated until it meets the equi-
vertices correspond to the random variables X1… Xn, as well librium state [60]. Penalty parameter (C) is chosen as 1,
as whose edges denote direct dependencies between the var- and s polynomial kernel is selected based on the trial-
iables. θ denotes the set of parameters which quantifies the and-error method for the SMO classification model,
network. For the BayesNet classification model, in the pres- because they provide better model performance.
ent study, a simple estimator is used, and the batch size is
chosen as 100 to get better classification accuracy. 3.7.3. Naïve Bayes Classifier. The Naïve Bayes (NB) classi-
fier is a probabilistic machine learning model based on
3.7.2. Sequential Minimal Optimization Classifier. Sequential Bayes theorem; that is, very useful for classifying a large
minimal optimization (SMO) is an important algorithm that amount of data. It anticipates based on the probability of
can solve the quadratic programming problem as well as an object. Thus, the model is developed, and the common
Table 3: Calculated values of ROC and PRC areas for test data itself, CVF 4, CVF 10, and PS 34% of test data using BayesNet, SMO, NB, and
RF classifiers.
ROC areas PRC areas

Testing options
BayesNet SMO NB RF BayesNet SMO NB RF
Test data itself 0.933 0.984 0.900 1.000 0.798 0.944 0.716 1.000
CVF 4 0.899 0.925 0.890 0.956 0.723 0.795 0.697 0.894
CVF 10 0.901 0.924 0.891 0.956 0.727 0.791 0.697 0.897
PS 34% 0.895 0.889 0.892 0.932 0.718 0.713 0.703 0.848
1 1
Normalized expected cost

0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Probability cost function Probability cost function
SMO RF SMO RF
(a) (b)
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 8: Cost curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for the test data itself using BayesNet, SMO,
NB, and RF classifiers.
input to the model is batch size (number of experiments dure consists of building simple decision trees and finally
carried out) [61], whose value is selected as 100 for better making the decision based on the votes. It looks for the
accuracy of the model. best feature in every step (splitting nodes and growing
trees), which predominantly results in a superior model.
3.7.4. Random Forest Classifier. The random forest (RF) Batch size, number of features, and seed value are the com-
classifier is a supervised machine learning algorithm that mon inputs in the RF model [61]. For better performance,
creates a forest with a bunch of trees. Its working proce- we have considered batch size as 100 and seed value as 1.
1 1

0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1

0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 9: Cost curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for CVF 4 of the test data using BayesNet,
3.8. Classified Result. In this research, it is expected that the the CVF 10 [19], and percentage split (PS) 34% [51] of
final classified results will be obtained as glioma, meningi- test data, by employing the Equations (1)–(11) to measure
oma, no-tumor, and pituitary tumors from MRI datasets. the performance of our proposed model.
Then, we wish that the qualitative and quantitative values
of the performance measurement of the proposed model’s TP + TN
will be attained to evaluate the model performance. Accuracy = , ð1Þ
TP + TN + FP + FN
3.9. Performance Measurement. A confusion matrix has

TP
been utilized to calculate the performance of a classifier. Sensitivity or Recall or True Positive Rate = , ð2Þ
The performances are described by four components of TP + FN
the confusion matrix, such as true positive (TP), true neg-
ative (TN), false positive (FP), and false negative (FN), TN
Specificity or True Negative Rate = , ð3Þ
which are whole in number and not in rate. The values TN + FP
of the performance parameters, such as accuracy, sensitiv-
ity, specificity, kappa statistics, false positive rate, preci- Observed Agreement − Expected Agreement
sion, f-measure, and Matthew’s correlation coefficient Kappa statistics = :
1 − Expected Agreement
(MCC), have been calculated for four testing options: the
test data itself, the cross-validation fold (CVF) 4 [16, 34], ð4Þ
1 1
0.8 0.8

0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 10: Cost curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for CVF 10 of the test data using BayesNet,
Here, 4 marginal, r1m is the row 1 marginal, r2m is the row 2 mar-
ginal, r3m is the row 3 marginal, r4m is the row 4 marginal,
Total Number of Correctly Classified and n is the total number of observations.
Observed Agreement = ,
n
ð5Þ FP
False Positive Rate ðFPRÞ = , ð7Þ
FP + TN
and
FN
Expected Agreement False Negative RateðFNRÞ = , ð8Þ
ÀÀ 1 Á Á ÀÀ Á Á ÀÀ Á Á ÀÀ Á Á FN + TP
cm × r1m /n + c2m × r2m /n + c3m × r3m /n + c4m × r4m /n
= ,
n TP
Precision = , ð9Þ
ð6Þ TP + FP
where c1m is the column 1 marginal, c2m is the column 2 Precision × Recall
F − measure = 2 × , ð10Þ
marginal, c3m is the column 3 marginal, c4m is the column Precision + Recall
1 1
0.8 0.8

0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(a) (b)
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SMO RF SMO RF
(c) (d)
Figure 11: Cost curves of (a) glioma, (b) meningioma, (c) no tumor, and (d) pituitary tumors for PS 34% of the test data using BayesNet,
that represents the performance of the classifier based on

Matthew’ s correlation coefficient ðMCCÞ the misclassification cost [63].
TP × TN − FP × FN ð11Þ
= pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi :
ðTP + FPÞðTP + FNÞðTN + FPÞðTN + FNÞ 4. Results and Discussion
4.1. Confusion Matrix. Table 2 represents the confusion
Different types of curves, such as the receiver operating matrix for the test data itself: CVF 4, CVF 10, and PS 34%
characteristics (ROC) curve, precision-recall curve (PRC), of the test data using BayesNet, SMO, NB, and RF classifiers,
and cost curve, have been plotted to understand the per- respectively. For the test data itself, the table shows that our
formance of every classifier. The ROC curve is utilized to models with the BayesNet classifier can classify 203 gliomas,
quantify how well the zone could be recognized from 175 meningiomas, 243 no-tumors, and 216 pituitary tumor
noise by plotting the true positive rate against the false images accurately. The SMO classifier can classify 252 glio-
positive rate [28, 32]. The PRC is a well-known model mas, 256 meningiomas, 269 no-tumors, and 263 pituitary
execution measurement to assess the binary classification images accurately. The NB classifier can classify 197 glioma
model [62]. A cost curve can be defined as a visual curve cases, 144 meningioma cases, 235 no-tumor cases, and 207
Table 4: Calculated values of different performance metrics for test data itself, CVF 4, CVF 10, and PS 34% of test data using BayesNet,
Test data itself CVF 4 of the test data

Performance metrics
BayesNet SMO NB RF BayesNet SMO NB RF
Accuracy (%) 88.750 98.148 86.250 100 85.972 92.500 85.741 91.435
Sensitivity (%) 77.500 96.296 72.500 100 71.946 85.000 71.482 82.871
Specificity (%) 92.500 98.763 90.834 100 90.648 95.000 90.494 94.290
FPR 0.075 0.012 0.092 0.000 0.094 0.050 0.095 0.057
FNR 0.225 0.037 0.275 0.000 0.281 0.150 0.286 0.171
Precision 0.775 0.963 0.722 1.000 0.722 0.850 0.710 0.826
F-measure 0.775 0.963 0.722 1.000 0.720 0.850 0.711 0.827
Kappa 0.700 0.950 0.633 1.000 0.626 0.800 0.620 0.772
MCC 0.700 0.951 0.632 1.000 0.627 0.800 0.618 0.770
CVF 10 of the test data PS 34% of the test data
BayesNet SMO NB RF Bayes net SMO NB RF
Accuracy (%) 85.834 92.454 85.324 91.389 85.484 89.200 85.204 89.271
Sensitivity (%) 71.667 84.907 70.648 82.778 71.102 78.534 70.557 78.680
Specificity (%) 90.556 94.969 90.216 94.259 90.324 92.790 90.135 92.851
FPR 0.094 0.050 0.098 0.057 0.097 0.072 0.099 0.071
FNR 0.283 0.151 0.294 0.172 0.289 0.215 0.295 0.214
Precision 0.717 0.849 0.702 0.824 0.709 0.783 0.697 0.783
F-measure 0.716 0.849 0.703 0.825 0.707 0.784 0.698 0.783
Kappa 0.622 0.799 0.609 0.770 0.613 0.712 0.606 0.714
MCC 0.622 0.799 0.607 0.769 0.612 0.711 0.602 0.713
pituitary cases accurately. The RF classifier, our model, can rately; while we have utilized SMO classifier, our model
classify all the classes accurately. can classify 132 glioma images, 116 meningioma images,
For CVF 4 of the test data, in the table, we can see that 161 no-tumors, and 150 pituitary images accurately; again,
our model with the BayesNet classifier can classify 181 gli- while we have utilized NB classifier, our model can classify
oma images, 158 meningioma images, 233 no-tumors, and 120 glioma images, 84 meningioma images, 154 no-
205 pituitary images accurately; with the SMO classifier, it tumors, and 144 pituitary images accurately; while we have
can classify 224 glioma images, 200 meningioma images, utilized RF classifier, our model can classify 134 glioma
251 no-tumors, and 243 pituitary images accurately; with images, 107 meningioma images, 163 no-tumors, and 156
the NB classifier, it can classify 191 glioma images, 138 pituitary images accurately.
meningioma images, 235 no-tumors, and 208 pituitary
images accurately; and with the RF classifier can classify 4.2. ROC and PRC Curves of Test Data. We have plotted
207 glioma images, 187 meningioma images, 256 no-tumors, ROC (Figures 4–7) and PRC (not shown in this paper)
and 245 pituitary images accurately. curves of glioma, meningioma, no-tumor, and pituitary
As shown in the table, for CVF 10 of the test data, while tumors for the test data itself, CVF 4, CVF 10, and PS 34%
we utilizing BayesNet classifier, our model can classify 183 of test data using BayesNet, SMO, NB, and RF classifiers,
glioma images, 153 meningioma images, 232 no-tumors, respectively. And then, we have calculated the values of the
and 206 pituitary images accurately; while utilizing SMO area under each ROC and PRC curves. The average values
classifier, our model can classify 211 glioma images, 208 are summarized in Table 3. As shown in the table, the aver-
meningioma images, 254 no-tumors, and 244 pituitary age values of the ROC area (i) for the test data itself are
images accurately; again, while utilizing NB classifier, our 0.933, 0.984, 0.900, and 1.000; (ii) for CVF 4 are 0.899,
model can classify 189 glioma images, 134 meningioma 0.925, 0.890, and 0.956; (iii) for CVF 10 are 0.901, 0.924,
images, 235 no-tumors, and 205 pituitary images accurately; 0.891, and 0.956; and (iv) for PS 34% are 0.895, 0.889,
while utilizing RF classifier, our model can classify 214 gli- 0.892, and 0.932 of the test data using BayesNet, SMO, NB,
oma images, 182 meningioma images, 256 no-tumors, and and RF classifiers, respectively. These values indicate that
242 pituitary images accurately. the RF classifier performs better than the other classifiers
In the case of PS 34% of the test data, the table dem- in every case. The table also similarly reveals that the RF
onstrates that while we have utilized BayesNet classifier, classifier gives the highest PRC area values of 1.000, 0.894,
our model can classify 116 glioma images, 98 meningioma 0.897, and 0.848 for the test data itself, CVF 4, CVF 10,
images, 152 no-tumors, and 140 pituitary images accu- and PS 34%, respectively.
Table 5: Comparison with other existing findings.
Ref. Classifiers Database Accuracy (%)

AlexNet 60.20
[15] BRATS2015
VGG16 98.20
[26] CNN Nanfang Hospital, General Hospital, China 97.3
[27] CNN Kaggle 97.2
DCNN 94.5
[28] M/s. Devaki Scan Centre, India
MDCNN 96.4
DenseNet+RNN 84.61
Nanfang Hospital, General Hospital,
[29] DenseNet+LSTM 92.13
Tianjin Medical University, China
DenseNet+DenseNet 86.68
[30] CNN BRATS 97.87
[32] GoogLeNET FIGSHARE 98
Nanfang Hospital and General Hospital,
96.13
[34] CNN Tianjin Medical University, China
REMBRANDT 98.7
[35] CNN FIGSHARE 98.51
[36] CNN BRATS2015, radiopedia 97.5
[64] CNN FIGSHARE 98.71
BRATS2012, BRATS2013,
[65] Deep wavelet auto-encoder 99.30
BRATS2014, BRATS2015
[66] Deep recurrent long short-term memory BRATS2015 99
[67] GoogLeNet+AlexNet TJU Hospital, China. 99.51
AlexNet CNN+BayesNet 88.75
AlexNet CNN+SMO 98.15
[proposed] Kaggle
AlexNet CNN+NB 86.25
AlexNet CNN+RF 100
4.3. Cost Curve Analysis of Test Data. We have also analyzed For the test data itself, Table 4 shows that the values of accuracy,
cost curves which yields model performance based on the sensitivity, specificity, FPR, FNR, precision, f-measure, kappa,
minimum misclassification cost. The minimum misclassifica- and MCC are higher for the RF classifier compared to those
tion cost can also be defined as the minimum area under the values of the other classifiers, exhibiting that the RF classifier
bottom of the envelope. Therefore, the minimum region below gives excellent performance with 100% accuracy.
the lower end of the range lowers the cost of misclassification For the case of CVF 4 of test data, Table 4 shows that the
and thus provides better classification performance. Figures 8– values of accuracy, sensitivity, specificity, precision, f-mea-
11 represent the cost curves for different testing options, such sure, kappa, and MCC are higher and the values of FPR
as the test data itself, CVF 4, CVF 10, and PS 34% of the test and FNR are lower for the SMO classifier than those values
data employing BayesNet, SMO, NB, and RF classifiers for gli- for the other classifiers, indicating that the SMO classifier
oma, meningioma, no-tumor, and pituitary classes, respec- gives better performance with 92.50% accuracy. Similar
tively. As seen in Figure 8, the RF classifier very clearly offers results are seen in the case of CVF 10 of test data, where
minimal misclassification cost; that is, better classification per- the SMO classifier gives 92.45% accuracy.
formance. However, the SMO classifier shows a slightly lower However, PS 34% of the test data in the table shows that
misclassification cost in Figures 9–11. the values of accuracy, sensitivity, specificity, FPR, FNR,
kappa, and MCC are higher for the RF classifier than the
4.4. Performance Evaluation. Finally, we have evaluated the other classifiers. The values of precision are the same for
performance of our models based on their respective confusion both SMO and RF classifiers, and the values of f-measure
matrices. We have calculated the values of accuracy, sensitivity, are very slightly lower for the RF classifier compared to the
specificity, FPR, FNR, precision, f-measure, kappa and MCC SMO classifier. From these statistical values, we can assume
and summarized them in Table 4, which represents the values that the RF classifier reveals a better performance with
for the test data itself: CVF 4, CVF 10, and PS 34%, respectively. 89.27% accuracy.
4.5. Comparison with Other Works. In this study, we have References

obtained the highest accuracy of 88.75%, 98.15%, 86.25%,
and 100% for AlexNet CNN with (i) BayesNet, (ii) NB, [1] D. N. Louis, A. Perry, G. Reifenberger et al., “The 2016
(iii) SMO, and (iv) RF classifiers using the test data itself, World Health Organization classification of tumors of the
respectively. Table 5 represents the comparison of our pro- central nervous system: a summary,” Acta Neuropathologica,
vol. 131, no. 6, pp. 803–820, 2016.
posed model with several contemporary findings in terms
of accuracy. As shown in the table, our model is highly [2] N. B. Bahadure, A. K. Ray, and H. P. Thethi, “Image analysis
appreciable, effective, and novel. for MRI based brain tumor detection and feature extraction
using biologically inspired BWT and SVM,” International
Journal of Biomedical Imaging, vol. 2017, Article ID 9749108,
12 pages, 2017.
5. Conclusion
[3] K. Popuri, D. Cobzas, A. Murtha, and M. Jägersand, “3D var-
In this study, a novel and effective technique for feature iational brain tumor segmentation using Dirichlet priors on a
extraction and classification of brain MRIs is presented clustered feature set,” International Journal of Computer
Assisted Radiology and Surgery, vol. 7, no. 4, pp. 493–506,
for classifying brain tumors into glioma, meningioma, no-
2012.
tumor, and pituitary classes. An AlexNet CNN with Bayes-
Net/SMO/NB/RF models is approached. Performances of [4] K. Fukushima, “Neocognitron: A self-organizing neural net-
work model for a mechanism of pattern recognition unaffected
the model are evaluated using the test dataset. We have
by shift in position,” Biological Cybernetics, vol. 36, no. 4,
calculated the values of performance parameters from pp. 193–202, 1980.
respective confusion matrix, namely, accuracy, sensitivity,
[5] M. J. Mckeown and D. A. Ramsay, “Classification of astrocy-
specificity, FPR, FNR, precision, f-measure, kappa statistics, tomas and malignant astrocytomas by principal components
MCC, ROC area, and PRC area for four testing options: Analysis and a neural net,” Journal of Neuropathology and
the test data itself, CVF 4, CVF 10, and PS 34% of test Experimental Neurology, vol. 55, no. 12, pp. 1238–1245,
data. We have also explained ROC, PRC, and cost curves 1996.
to measure the classification performance. We have [6] S. W. Coons, P. C. Johnson, B. W. Scheithauer, A. J. Yates, and
acquired 100%, 98.15%, 88.75%, and 86.25% of accuracy D. K. Pearl, “Improving diagnostic accuracy and interobserver
using the AlexNet CNN+RF, AlexNet CNN+SMO, AlexNet concordance in the classification and grading of primary glio-
CNN+BayesNet, and AlexNet CNN + NB models, respec- mas,” American Cancer Society, vol. 79, no. 7, pp. 1381–1393,
tively. This study presents a paramount technique for the 1997.
classification of brain tumors from MRIs. [7] N. Belacela and M. R. Boulassel, “Multicriteria fuzzy assign-
However, the proposed model was evaluated using a ment method: a useful tool to assist medical diagnosis,” Artifi-
moderately sized dataset, which is one of the limitations of cial Intelligence in Medicine, vol. 21, no. 1-3, pp. 201–207,
the current work. So, it is essential to evaluate the model 2001.
with big data size in the future to see how well it performs. [8] H. Gemmeke and N. V. Ruiter, “3D ultrasound computer
Another limitation is that the proposed approach has not tomography for medical imaging,” Nuclear Instruments and
been tested employing the MRIs of real-patient in Bangla- Methods in Physics Research, vol. 580, no. 2, pp. 1057–1065,
desh. In the future, we will try to deploy this work on a 2007.
real-time medical diagnostic system by acquiring the MRIs [9] E. S. A. El-Dahshan, H. M. Mohsen, K. Revett, and A. B. M.
from different hospitals and diagnostic centers in Bangla- Salem, “Computer-aided diagnosis of human brain tumor
desh. We believe that our approach is very much applicable through MRI: A survey and a new algorithm,” Expert Systems
in the health sector, offering innovative solutions with high with Applications, vol. 41, no. 11, pp. 5526–5545, 2014.
precision for medical imaging, especially for diagnosing [10] B. Dogan, O. Demir, and S. K. Calik, “Computer-aided detec-
brain tumors more accurately. tion of brain tumors using morphological reconstruction,”
Uludag University Journal of The Faculty of Engineering,
vol. 21, no. 2, pp. 257–268, 2016.
Data Availability [11] Y. Yan, X. J. Yao, S. H. Wang, and Y. D. Zhang, “A Survey of
computer-aided tumor diagnosis based on convolutional neu-
Experimental codes and the data of this work are available ral network,” Biology, vol. 10, no. 11, pp. 1–44, 2021.
from the corresponding author upon request at any time. [12] A. Rehman, M. A. Khan, T. Saba, Z. Mehmood, U. Tariq, and
N. Ayesha, “Microscopic brain tumor detection and classifica-
tion using 3D CNN and feature selection architecture,”
Conflicts of Interest Microscopy Research and Technique, vol. 84, no. 1, pp. 133–
149, 2021.
The authors declare no conflicts of interest. [13] J. Amin, M. Sharif, A. Haldorai, M. Yasmin, and R. S. Nayak,
“Brain tumor detection and classification using machine learn-
ing: a comprehensive survey,” Complex & Intelligent Systems,
Acknowledgments vol. 8, no. 4, pp. 3161–3183, 2022.
[14] A. A. Akinyelu, F. Zaccagna, J. T. Grist, M. Castelli, and
We wish to express our gratitude to Khulna University, L. Rundo, “Brain tumor diagnosis using machine learning,
Khulna, Bangladesh-9208. convolutional neural networks, capsule neural networks and
vision transformers, applied to MRI: a survey,” Journal Imag- [29] Y. Zhou, Z. Li, H. Zhu et al., “Holistic brain tumor screening
ing, vol. 8, no. 8, pp. 205–240, 2022. and classification based on DenseNet and recurrent neural
[15] N. A. Ali, A. R. Syafeeza, L. J. Geok, Y. C. Wong, N. A. Hamid, network,” in Brainlesion: Glioma, Multiple Sclerosis, Stroke
and A. S. Jaafar, “Design of automated computer-aided classi- and Traumatic Brain Injuries. BrainLes 2018. Lecture Notes
fication of brain tumor using deep learning,” in Intelligent and in Computer Science, vol 11383Springer, Cham.
Interactive Computing. Lecture Notes in Networks and Systems, [30] T. Hossain, F. S. Shishir, M. Ashraf, M. A. Al Nasim, and F. M.
vol 67, V. Piuri, V. Balas, S. Borah, and S. Syed Ahmad, Eds., Shah, “Brain tumor detection using convolutional neural net-
Springer, Singapore, 2019. work,” in 2019 1st International Conference on Advances in
[16] B. S. Kumar, “Identification and classification of brain tumor Science, Engineering and Robotics Technology (ICASERT),
images using efficient classifier,” International Journal of Engi- pp. 1–6, Dhaka, Bangladesh, 2019.
neering and Advanced Technology, vol. 8, no. 6, pp. 3677–3683, [31] P. M. Krishnammal and S. S. Raja, “Convolutional neural net-
2019. work based image classification and detection of abnormalities
[17] T. Ateeq, M. N. Majeed, S. M. Anwar et al., “Ensemble-classi- in MRI brain images,” in 2019 International Conference on
fiers-assisted detection of cerebral microbleeds in brain MRI,” Communication and Signal Processing (ICCSP), pp. 0548–
Computers and Electrical Engineering, vol. 69, pp. 768–781, 0553, Chennai, India, 2019.
2018. [32] S. Deepak and P. M. Ameer, “Brain tumor classification using
[18] J. Cheng, W. Huang, S. Cao et al., “Enhanced performance of deep CNN features via transfer learning,” Computers in Biol-
brain tumor classification via tumor region augmentation ogy and Medicine, vol. 111, article 103345, pp. 1–12, 2019.
and partition,” PLoS One, vol. 10, no. 10, pp. 1–13, 2015. [33] E. I. Papageorgiou, P. P. Spyridonos, D. T. Glotsos et al., “Brain
[19] D. R. Patel, H. Thakker, M. B. Kiran, and V. Vakhariya, “Sur- tumor characterization using the soft computing technique of
face roughness prediction of machined components using gray fuzzy cognitive maps,” Applied Soft Computing, vol. 8, no. 1,
level co-occurrence matrix and bagging tree,” FME Transac- pp. 820–828, 2008.
tions, vol. 48, no. 2, pp. 468–475, 2020. [34] H. H. Sultan, N. M. Salem, and W. Al-Atabany, “Multi-classi-
[20] F. A. Zadeh, M. V. Ardalani, A. R. Salehi, R. J. Farahani, fication of brain tumor images using deep neural network,”
M. Hashemi, and A. H. Mohammed, “An analysis of new fea- IEEE Access, vol. 7, pp. 69215–69225, 2019.
ture extraction methods based on machine learning methods [35] N. Abiwinanda, M. Hanif, S. T. Hesaputra, A. Handayani, and
for classification radiological images,” Computational Intelli- T. R. Mengko, “Brain tumor classification using convolutional
gence and Neuroscience, vol. 2022, Article ID 3035426, 13 neural network,” in World Congress on Medical Physics and
pages, 2022. Biomedical Engineering 2018, pp. 183–189, Springer, 2019.
[21] G. T. Reddy, S. Bhattacharya, and S. S. Ramakrishnan, “An [36] J. Seetha and S. S. Raja, “Brain tumor classification using con-
ensemble based machine learning model for diabetic retinopa- volutional neural networks,” Biomedical & Pharmacology
thy classification,” in 2020 International Conference on Emerg- Journal, vol. 11, no. 3, pp. 1457–1461, 2018.
ing Trends in Information Technology and Engineering, pp. 1– [37] H. Selvaraj, S. T. Selvi, D. Selvathi, and L. Gewali, “Brain MRI
6, Vellore, India, 2020. slices classification using least squares support vector
[22] R. H. Jhaveri, A. Revathi, K. Ramana, R. Raut, and R. K. Dha- machine,” International Journal of Intelligent Computing in
naraj, “A review on machine learning strategies for real-world Medical Sciences & Image Processing, vol. 1, no. 1, pp. 21–33,
engineering applications,” Mobile Information Systems, 2007.
vol. 2022, Article ID 1833507, 26 pages, 2022. [38] A. Sarkar, M. Maniruzzaman, M. S. Ahsan, M. Ahmad,
[23] G. T. Reddy, M. P. K. Reddy, K. Lakshmanna, D. S. Rajput, M. I. Kadir, and S. T. Islam, “Identification and classifica-
R. Kaluri, and G. Srivastava, “Hybrid genetic algorithm and a tion of brain tumor from MRI with feature extraction by
fuzzy logic classifier for heart disease diagnosis,” Evolutionary support vector machine,” in 2020 International Conference
Intelligence, vol. 13, no. 2, pp. 185–196, 2020. for Emerging Technology (INCET), pp. 1–4, Belgaum, India,
[24] U. Ahmed, R. H. Jhaveri, G. Srivastava, and J. C. W. Lin, 2020.
Explainable deep attention active learning for sentimental ana- [39] Z. Ullah, M. U. Farooq, S. H. Lee, and D. An, “A hybrid
lytics of mental disorder, Association for Computing Machin- image enhancement based brain MRI images classification
ery, 2022. technique,” Medical Hypotheses, vol. 143, article 109922,
[25] P. John, “Brain tumor classification using wavelet and texture 2020.
based neural network,” International Journal of Scientific and [40] A. Kharrat, K. Gasmi, M. B. Messaoud, N. Benamrane, and
Engineering Research, vol. 3, no. 10, pp. 1–7, 2012. M. Abid, “A hybrid approach for automatic classification of
[26] F. J. Díaz-Pernas, M. Martínez-Zarzuela, M. Antón-Rodríguez, brain MRI using genetic algorithm and support vector
and D. González-Ortega, “A deep learning approach for brain machine,” Leonardo Journal of Sciences, vol. 17, no. 1,
tumor classification and segmentation using a multiscale convo- pp. 71–82, 2010.
lutional neural network,” Healthcare, vol. 9, no. 2, p. 153, 2021. [41] N. V. Shree and T. N. R. Kumar, “Identification and classifica-
[27] A. Çinar and M. Yildirim, “Detection of tumors on brain MRI tion of brain tumor MRI images with feature extraction using
images using the hybrid convolutional neural network archi- DWT and probabilistic neural network,” Brain Informatics,
tecture,” Medical Hypotheses, vol. 139, pp. 109684–109688, vol. 5, no. 1, pp. 23–30, 2018.
2020. [42] M. Arunachalam and S. Royappan Savarimuthu, “An efficient
[28] D. J. Hemanth, J. Anitha, A. Naaji, O. Geman, D. E. A. and automatic glioblastoma brain tumor detection using shift-
Popescu, and L. Hoang Son, “A Modified deep convolutional invariant shearlet transform and neural networks,” Interna-
neural network for abnormal brain image classification,” IEEE tional Journal of Imaging Systems and Technology, vol. 27,
Access, vol. 7, pp. 4275–4283, 2019. no. 3, pp. 216–226, 2017.
[43] S. Sajid, S. Hussain, and A. Sarwar, “Brain tumor detection and national Journal of Applied Engineering Research, vol. 55,
segmentation in MR images using deep learning,” Arabian pp. 1045–1049, 2015.
Journal for Science and Engineering, vol. 44, no. 11, [61] V. Vakharia and R. Gujar, “Prediction of compressive strength
pp. 9249–9261, 2019. and portland cement composition using cross-validation and
[44] P. Saxena, A. Maheshwari, and S. Maheshwari, “Predictive feature ranking techniques,” Construction and Building Mate-
modeling of brain tumor: a deep learning approach,” in Inno- rials, vol. 225, pp. 292–301, 2019.
vations in Computational Intelligence and Computer Vision. [62] T. Saito and M. Rehmsmeier, “The precision-recall plot is
Advances in Intelligent Systems and Computing, vol more informative than the ROC plot when evaluating binary
1189Springer, Singapore. classifiers on imbalanced datasets,” PLoS One, vol. 10, no. 3,
[45] S. Khawaldeh, U. Pervaiz, A. Rafiq, and R. S. Alkhawaldeh, pp. 1–21, 2015.
“Noninvasive grading of glioma tumor using magnetic reso- [63] C. Drummond and R. C. Holte, “Cost curves: An improved
nance imaging with convolutional neural networks,” Applied method for visualizing classifier performance,” Machine
Sciences, vol. 8, no. 1, p. 27, 2018. Learning, vol. 65, no. 1, pp. 95–130, 2006.
[46] S. Preethi and P. Aishwarya, “Combining wavelet texture fea- [64] R. Chelghoum, A. Ikhlef, A. Hameurlaine, and S. Jacquir,
tures and deep neural network for tumor detection and seg- “Transfer learning using convolutional neural network archi-
mentation over MRI,” Journal of Intelligent Systems, vol. 28, tectures for brain tumor classification from MRI images,” in
no. 4, pp. 571–588, 2019. Artificial Intelligence Applications and Innovations. AIAI
[47] M. A. Khan, I. Ashraf, M. Alhaisoni et al., “Multimodal brain 2020. IFIP Advances in Information and Communication
tumor classification using deep learning and robust feature Technology, vol 583Springer, Cham.
selection: a machine learning application for radiologists,” [65] I. A. E. Kader, G. Xu, Z. Shuai et al., “Brain tumor detection
Diagnostics, vol. 10, no. 8, p. 565, 2020. and classification on MR images by a deep wavelet auto-
[48] M. Siar and M. Teshnehlab, “Brain tumor detection using deep encoder model,” Diagnostics, vol. 11, no. 9, pp. 1–19, 2021.
neural network and machine learning algorithm,” in 2019 9th [66] E. Aarthi, S. Jana, W. G. Theresa et al., “Detection and classifi-
International Conference on Computer and Knowledge Engi- cation of MRI brain tumors using S3-DRLSTM based deep
neering (ICCKE), pp. 363–368, Mashhad, Iran, 2019. learning model,” International Journal of Electrical and Elec-
[49] M. Li, L. Kuang, S. Xu, and Z. Sha, “Brain tumor detection tronics Research, vol. 10, no. 3, pp. 597–603, 2022.
based on multimodal information fusion and convolutional [67] N. A. Samee, N. F. Mahmoud, G. Atteia et al., “Classification
neural network,” IEEE Access, vol. 7, pp. 180134–180146, framework for medical diagnosis of brain tumor with an effec-
2019. tive hybrid transfer learning model,” Diagnostis, vol. 12, no. 10,
[50] MatLab, Version 9.2 (R2017a), The MathWorks Inc., Natick, p. 2541, 2022.
Massachusetts, 2017.
[51] E. Frank, M. A. Hall, and I. H. Witten, The WEKA workbench.
Online appendix for “Data mining: practical machine learning
tools and techniques”, Morgan Kaufmann, Fourth Edition edi-
tion, 2016.
[52] https://www.kaggle.com/datasets/sartajbhuvaji/brain-tumor-
classification-mri.
[53] https://www.kaggle.com/datasets/ahmedhamada0/brain-
tumor-detection?-Resource=download.
[54] M. Coşkun, Ö. Yildirim, U. Ç. A. R. Ayşegül, and Y. Demir,
“An overview of popular deep learning methods,” European
Journal of Technique, vol. 7, no. 2, pp. 165–176, 2017.
[55] https://neurohive.io/en/popular-networks/alexnet-imagenet-
classification-with-deep-convolutional-neural-networks/.
[56] K. Simonyan and A. Zisserman, “Very deep convolutional net-
works for large- scale image recognition,” 2014, http://arxiv
.org/abs/1409.1556.
[57] X. Han, Y. Zhong, L. Cao, and L. Zhang, “Pre-trained Alexnet
architecture with pyramid pooling and supervision for high
spatial resolution remote sensing image scene classification,”
Remote Sensing, vol. 9, no. 8, p. 848, 2017.
[58] A. M. Carvalho, T. T. Roos, A. L. Oliveira, and P. Myllymäki,
“Discriminative learning of Bayesian networks via factorized
conditional log-likelihood,” Journal of Machine Learning
Research, vol. 12, pp. 2181–2210, 2011.
[59] N. Friedman, D. Geiger, and M. Goldszmidt, “Bayesian net-
work classifiers,” Machine Learning, vol. 29, no. 2/3, pp. 131–
163, 1997.
[60] S. P. Rajamohana, K. Umamaheswari, and R. Karthiga, “Senti-
ment classification based on LDA using SMO classifier,” Inter-

An Effective and Novel Approach For Brain Tumor CL

Uploaded by

Copyright:

Available Formats

An Effective and Novel Approach For Brain Tumor CL

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

An Effective and Novel Approach For Brain Tumor CL

Uploaded by

Copyright:

Available Formats

Hindawi

Alok Sarkar ,1 Md. Maniruzzaman ,1 Mohammad Ashik Alahe ,1

Correspondence should be addressed to Alok Sarkar; [email protected]

Academic Editor: Akhilesh Pathak

1. Introduction structure of the brain to diagnose tumors. The most popular

Training image set Test image set

Feature extraction Feature extraction

Extracted features Extracted features

Glioma Meningioma No-tumor Pituitary

accuracy was obtained. Updated versions of CNN architec-

Fully connected layer

Figure 2: Basic CNN architecture.

1st CON 11×11, 55×55×96

6×6×256 Max pool 3×3, 5th CON 3×3,

Figure 3: The proposed architecture of AlexNet.

BayesNet classiﬁer SMO classiﬁer

Glioma 203 36 6 25 Glioma 252 17 1 0

BayesNet classiﬁer SMO classiﬁer

BayesNet classiﬁer SMO classiﬁer

Pituitary 20 41 3 206 Pituitary 9 15 2 244

BayesNet classiﬁer SMO classiﬁer

True positive rate

True positive rate

False positive rate False positive rate

True positive rate

True positive rate

True positive rate

True positive rate

True positive rate

True positive rate

3.7. Classiﬁers be materialized by training the SVMs. At this issue, the

ROC areas PRC areas

Normalized expected cost

Normalized expected cost

Normalized expected cost

Normalized expected cost

Normalized expected cost

3.9. Performance Measurement. A confusion matrix has

Normalized expected cost

Normalized expected cost

Normalized expected cost

Probability cost function Probability cost function

Normalized expected cost

Probability cost function Probability cost function

that represents the performance of the classiﬁer based on

Test data itself CVF 4 of the test data

Table 5: Comparison with other existing ﬁndings.

Ref. Classiﬁers Database Accuracy (%)

4.5. Comparison with Other Works. In this study, we have References

You might also like