Improved Arabic Alphabet Characters Classification Using.pdf

Research Article
Improved Arabic Alphabet Characters Classification Using
Convolutional Neural Networks (CNN)
Nesrine Wagaa ,1,2,3
Hichem Kallel ,1,2,3
and Nédra Mellouli 1,2,3
1
National Institute of Applied Sciences and Technology (INSAT) at University of Carthage, LARATSI Laboratory, Cedex 1080,
Tunis, Tunisia
2
MedTech at South Mediterranean University, Cedex 1053, Tunis, Tunisia
3
Laboratory of Advanced Computer Science at Paris 8 University, LIASD (EA4383), France
Correspondence should be addressed to Nesrine Wagaa; nesrinewagah@gmail.com
Received 19 March 2021; Revised 4 November 2021; Accepted 16 December 2021; Published 11 January 2022
Academic Editor: Paolo Gastaldo
Copyright © 2022 Nesrine Wagaa 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.
Handwritten characters recognition is a challenging research topic. A lot of works have been present to recognize letters of
different languages. The availability of Arabic handwritten characters databases is limited. Motivated by this topic of research, we
propose a convolution neural network for the classification of Arabic handwritten letters. Also, seven optimization algorithms are
performed, and the best algorithm is reported. Faced with few available Arabic handwritten datasets, various data augmentation
techniques are implemented to improve the robustness needed for the convolution neural network model. The proposed model is
improved by using the dropout regularization method to avoid data overfitting problems. Moreover, suitable change is presented
in the choice of optimization algorithms and data augmentation approaches to achieve a good performance. The model has been
trained on two Arabic handwritten characters datasets AHCD and Hijja. The proposed algorithm achieved high recognition
accuracy of 98.48% and 91.24% on AHCD and Hijja, respectively, outperforming other state-of-the-art models.
1. Introduction
Approximately a quarter of a billion people around the
world speak and write the Arabic language [1]. There are a lot
of historical books and documents that represent a crucial
data set for most Arabic countries written in the Arabic
language [1, 2].
Recently, the area of Arabic handwritten characters
recognition (AHCR) has received increased research at-
tention [3–5]. It is a challenging topic of computer vision
and pattern recognition [1]. This is due to the following:
(i) The difference between handwriting patterns [3].
(ii) The form similarity between Arabic alphabets [1, 3].
(iii) The diacritics of Arabic characters [6].
(iv) As shown in Figure 1, in the Arabic language the
shape of each handwritten character depends on its
position in the world. For example, here in the word
“‫”ءارمأ‬ the character “Alif” is written in two different
forms “‫”أ‬ and “‫”ا‬, where, in the Arabic language,
each character has between two and four shapes.
Table 1 shows the different shapes of the twenty-
eight Arabic alphabets.
With the development of deep learning (DL), convo-
lution neural networks (CNNs) have shown a significant
capability to recognize handwritten characters of different
languages [3, 7, 8]: Latin [9, 10], Chine [11], Devanagari [12],
Malayalam [11], etc.
Most researchers improved the CNN architecture to
achieve good handwritten characters recognition per-
formance [6, 13]. However, a neural network with ex-
cellent performance usually requires a good tuning of
CNN hyperparameters and a good choice of applied
optimization algorithms [14–16]. Also, a large amount of
training dataset [17, 18] is required to achieve outstanding
performance.
The main contributions of this research can be sum-
marized as follows:
Hindawi
Computational Intelligence and Neuroscience
Volume 2022,Article ID 9965426, 16 pages
https://doi.org/10.1155/2022/9965426

(i) Suggesting a CNN model for recognizing Arabic
handwritten characters.
(ii) Tuning of different hyperparameters to improve the
model performance.
(iii) Applying different optimization algorithms.
Reporting the effectiveness of the best ones.
(iv) Presenting different data augmentation techniques.
Reporting the influence of each method on the
improvement of Arabic handwritten characters
recognition.
(v) Mixing two different Arabic handwritten characters
datasets for shape varying. Testing the impact of the
presented data augmentation approaches on the
mixed dataset.
The rest of this paper is organized as follows. In Section 2,
we expose the related works in Arabic handwritten character
classification. In Sections 3 and 4, we describe the convolution
neural network architecture and the model tuning hyper-
parameters. In Section 5, we make a detailed description of
various used optimization algorithms. In Section 6, we de-
scribe the different utilized data augmentation techniques
chosen in this study. In Section 7, we provide an overview of
the experimental results showing the CNN distinguished
performance. Section 8 is conclusion and possible future
research directions.
2. Related Work
In recent years, many studies have addressed the classifi-
cation and recognition of letters, including Arabic hand-
written characters. On the other hand, there are a smaller
number of proposed approaches for recognizing individual
characters in the Arabic language. As a result, Arabic
handwritten character recognition is less common com-
pared to English, French, Chinese, Devanagari, Hangul,
Malayalam, etc.
Impressive results were achieved in the classification of
handwritten characters from different languages, using deep
learning models and in particular the CNN.
El-Sawy et al. [6] gathered their own Arabic Handwritten
Character dataset (AHCD) from 60 participants. AHCD
consists of 16.800 characters. They have achieved a classi-
fication accuracy of 88% by using a CNN model consisting of
2 convolutional layers. To improve the CNN performance,
regularization and different optimization techniques have
been implemented to the model. The testing accuracy was
improved to 94.93%.
Altwaijry and Turaiki [13] presented a new Arabic
handwritten letters dataset (named “Hijja”). It comprised
47.434 characters written by 591 participants. Their pro-
posed CNN model was able to achieve 88% and 97% testing
accuracy, using the Hijja and AHCD datasets, respectively.
Younis [19] designed a CNN model to recognize Arabic
handwritten characters. The CNN consisted of three con-
volutional layers followed by one final fully connected layer.
The model achieved an accuracy of 94.7% for the AHCD
database and 94.8% for the AIA9K (Arabic alphabet’s
dataset).
Latif et al. [20] designed a CNN to recognize a mix of
handwriting of multiple languages: Persian, Devanagari,
Eastern Arabic, Urdu, and Western Arabic. The input image
is of size (28 × 28) pixels, followed by two convolutional
layers, and then a max-pooling operation is applied to both
convolution layers. The overall accuracy of the combined
multilanguage database was 99.26%. The average accuracy is
around 99% for each individual language.
Alrobah and Albahl [21] analyzed the Hijja dataset and
found irregularities, such as some distorted letters, blurred
symbols, and some blurry characters. They used the CNN
model to extract the important features and SVM model for
data classification. They achieved a testing accuracy of 96.3%.
Mudhsh et al. [22] designed the VGG net architecture for
recognizing Arabic handwritten characters and digits. The
model consists of 13 convolutional layers, 2 max-pooling
layers, and 3 fully connected layers. Data augmentation and
Figure 1: One possible connected component box for the word
“‫”ءارمأ‬.
Table 1: Twenty-eight shapes of Arabic alphabets.
No. Name Isolated Beginning Middle End
1 Alif ‫ا‬ ‫ا‬ ‫إـ‬ ‫أـ‬
2 Baa ‫ب‬ ‫ـب‬ ‫ـبـ‬ ‫بـ‬
3 Tea ‫ت‬ ‫ـت‬ ‫ـتـ‬ ‫تـ‬
4 Thea ‫ث‬ ‫ـث‬ ‫ـثـ‬ ‫ثـ‬
5 Jam ‫ج‬ ‫ـج‬ ‫ـجـ‬ ‫جـ‬
6 Haa ‫ح‬ ‫ـح‬ ‫ـحـ‬ ‫حـ‬
7 Khaa ‫خ‬ ‫ـخ‬ ‫ـخـ‬ ‫خـ‬
8 Daal ‫د‬ ‫د‬ ‫دـ‬ ‫دـ‬
9 Thaal ‫ذ‬ ‫ذ‬ ‫ذـ‬ ‫ذـ‬
10 Raa ‫ر‬ ‫ر‬ ‫رـ‬ ‫رـ‬
11 Zaay ‫ز‬ ‫ز‬ ‫زـ‬ ‫زـ‬
12 Seen ‫س‬ ‫ـس‬ ‫ـسـ‬ ‫سـ‬
13 Sheen ‫ش‬ ‫ـش‬ ‫ـشـ‬ ‫شـ‬
14 Sad ‫ص‬ ‫ـص‬ ‫ـصـ‬ ‫صـ‬
15 Dhad ‫ض‬ ‫ـض‬ ‫ـضـ‬ ‫ضـ‬
16 Tah ‫ط‬ ‫ـط‬ ‫ـطـ‬ ‫طـ‬
17 Dha ‫ظ‬ ‫ـظ‬ ‫ـظـ‬ ‫ظـ‬
18 Ain ‫ع‬ ‫ـع‬ ‫ـعـ‬ ‫عـ‬
19 Ghen ‫غ‬ ‫ـغ‬ ‫ـغـ‬ ‫غـ‬
20 Fa ‫ف‬ ‫ـف‬ ‫ـفـ‬ ‫فـ‬
21 Qaf ‫ق‬ ‫ـق‬ ‫ـقـ‬ ‫قـ‬
22 Kaf ‫ك‬ ‫ـك‬ ‫ـكـ‬ ‫كـ‬
23 Lam ‫ل‬ ‫ـل‬ ‫ـلـ‬ ‫لـ‬
24 Meem ‫م‬ ‫ـم‬ ‫ـمـ‬ ‫مـ‬
25 Noon ‫ن‬ ‫ـن‬ ‫ـنـ‬ ‫نـ‬
26 Ha ‫ه‬ ‫ـه‬ ‫ـهـ‬ ‫هـ‬
27 Waw ‫و‬ ‫و‬ ‫وـ‬ ‫وـ‬
28 Yaa ‫ي‬ ‫ـي‬ ‫ـيـ‬ ‫يـ‬
2 Computational Intelligence and Neuroscience

dropout methods were used to avoid the overfitting prob-
lem. The model was trained and evaluated by using two
different datasets: the ADBase for the Arabic handwritten
digits classification topic and HACDB for the Arabic
handwritten characters classification task. The model
achieved an accuracy of 99.66% and 97.32% for ADBase and
HACDB, respectively.
Boufenar et al. [23] used the popular CNN architecture
Alexnet. It consists of 5 convolutional layers, 3 max-pooling
layers, and 3 fully connected layers. Experiments were
conducted on two different databases, OIHACDB-40 and
AHCD. Based on the good tuning of the CNN hyper-
parameters and by using dropout and minibatch techniques,
a CNN accuracy of 100% and 99.98% for OIHACDB-40 and
AHCD was achieved.
Mustapha et al. [24] proposed a Conditional Deep
Convolutional Generative Adversarial Network (CDCGAN)
for a guided generation of isolated handwritten Arabic
characters. The CDCGAN was trained on the AHCD dataset.
They achieved a 10% performance gap between real and
generated handwritten Arabic characters.
Table 2 summarizes the literature reviewed for recog-
nizing Arabic handwriting characters using the CNN
models. From the previous literature, we notice that most
CNN architectures have been trained by using adult Arabic
handwriting letters “AHCD”. In addition, we observe that
most researchers try to improve the performance through
the good tuning of the CNN model hyperparameters.
3. TheProposedArabicHandwrittenCharacters
Recognition System
As shown in Figure 2, the model that we proposed in this
study is composed of three principal components: CNN
proposed architecture, optimization algorithms, and data
augmentation techniques.
In this paper, the proposed CNN model contains four
convolution layers, two max-pooling operations, and an
ANN model with three fully hidden layers used for the
classification. To avoid the overfitting problems and improve
the model performance, various optimization techniques
were used such as dropout, minipatch, choice of the acti-
vation function, etc.
Figure 3 describes the proposed CNN model. Also, in
this work, the recognition performance of Arabic hand-
written letters was improved through the good choice of the
optimization algorithm and by using different data aug-
mentation techniques “geometric transformations, feature
space augmentation, noise injection, and mixing images.”
4. Convolution Neural Network Architecture
A CNN model [25–34] is a series of convolution layers
followed by fully connected layers. Convolution layers allow
the extraction of important features from the input data.
Fully connected layers are used for the classification of data.
The CNN input is the image to be classified; the output
corresponds to the predicted class of the Arabic handwritten
character.
4.1. Input Data. The input data is an image I of size
(m × m × s). (m × m) Defines the width and the height of the
image and s denotes the space or number of channels. The value
of s is1 for a grayscale image and equals 3 for a RGB color image.
4.2. Convolution Layer. The convolution layer consists of a
convolution operation followed by a pooling operation.
4.2.1. Convolution Operation. The basic concept of the
classical convolution operation between an input image I of
dimension (m × m) and a filter F of size (n × n) is defined as
follows (see Figure 4):
C � I ⊗ F. (1)
Here, ⊗ denotes the convolution operation. C is the con-
volution map of size (a× a), where a � (m − n + 2p/sL) + 1. sL
is the stride and denotes the number of pixels by which
F is sliding over I. p is the padding; often it is necessary to add a
bounding of zeros around I to preserve complete image in-
formation. Figure 4 is an example of the convolution operation
between an input image of dimension (8× 8) and a filter F of
size (3× 3). Here, the convolution map C is of size (6× 6) with a
stride sL � 1 and a padding p � 0.
Generally, a nonlinear activation function is applied on
the convolution map C. The commonly used activation
functions are Sigmoid [34–36], Hyperbolic Tangent “Tanh”
[35, 37], and Rectified Linear Unit “ReLU” [37, 38] where
Ca � f(C), (2)
here, Ca is the convolution map after applying the nonlinear
activation function f. Figure 5 shows the Ca map when the
ReLU activation function is applied on C.
4.2.2. Pooling Operation. The pooling operation is used to
reduce the dimension of Ca thus reducing the computational
complexity of the network. During the pooling operation, a
kernel K of size (sp × sp) is sliding over Ca. sp denotes the
number of patches by which K is sliding over Ca. In our
analysis sp is set to 2. The pooling operation is expressed as
P � pool Ca􏼁, (3)
where P is the pooling map and pool is the pooling oper-
ation. The commonly used pooling operations are average-
pooling, max-pooling, and min-pooling. Figure 6 describes
the concept of average-pooling and max-pooling operations
using a kernel of size (2 × 2) and a stride of 2.
4.3. Concatenation Operation. The concatenation operation
maps the set of the convoluted images into a vector called the
concatenation vector Y.
Y �
Pc
1
Pc
2
⋮
Pc
n
⎡
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣
⎤
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦
, (4)
Computational Intelligence and Neuroscience 3

here, Pc
i is the output of the ith
convolution layer. n denotes
the number of filters applied on the convoluted images Pc−1
i−1 .
4.4. Fully Connected Layer. The CNN classification opera-
tion is performed through the fully connected layer [39]. Its
input is the concatenation vector Y; the predicted class y is
the output of the CNN classifier. The classification oper-
ation is performed through a series of t fully connected
hidden layers. Each fully connected hidden layer is a
parallel collection of artificial neurons. Like synapses in the
biological brain, the artificial neurons are connected
through weights W. The model output of the ith
fully
connected hidden layer is
Table 2: Summary of Arabic handwritten characters recognition using CNN model.
References Year Dataset Type (size) Method Optimization
Accuracy
(%)
Loss
(%)
El-Sawy et al. [6] 2017 AHCD
Chars
(16,800)
CNN (i) Minibatch 94.93 5.1
Mudhsh et al. [22] 2017
ADBase
Digits
(6.600) CNN (based on VGG
net)
(ii) Dropout 99.6 —
HACDB
Chars
(70.000)
(iii) Data
augmentation
97.32 —
Boufenar et al. [23] 2017
OIHACDB Chars
(6.600)
CNN (based on
Alexnet)
(i) Dropout 100
—
AHCD (ii) Minibatch 99.98
Younis [19] 2018
AHCD Chars
(8.737)
CNN —
97.7 —
AIA9K 94.8 —
Latif et al. [20] 2018
Mix of handwriting of
multiple languages
Chars CNN — 99.26 0.02
Altwaijry and
Turaiki [13]
2020
Hijja Chars
(47,434)
CNN —
88 —
AHCD 97 —
Alrobah &Albahl
[21]
2021 Hijja
Chars
(47,434)
CNN + SVM — 96.3 —
Mustapha et al. [24] 2021 AHCD CDCGAN — — —
Input: Arabic handwritten character
image
Dropout Mini Batch
ELU cross-
entropy
Learning
Rate
SoftMax
AdaDelta
Nadam
Mixing images datasets
AHCD & Hijja
Speckle
Recognized character
Salt-and-pepper
Gaussian
Zooming
Flipping
Rotation
Geometric transformations &
Feature space augmentation
Data
augmentation
techniques
Shifting
Noise injection
AdaMax
Adagrad
ADAM
Gradient descent
with Momentum
RMSprop
Gradient descent
ANN Classifier
CNN
proposed
architecture
Optimization
Algorithms
Figure 2: Proposed recognition schema for Arabic handwritten characters datasets.

Yi
� f Hi
􏼐 􏼑, (5)
where the weight sum vector Hi
is
Hi
� Wi
Y(i− 1)
+ Bi
, (6)
here, f is a nonlinear activation function (sigmoid, Tanh,
ReLU, etc.). The bias value Bi
defines the activation level of
the artificial neurons.
5. CNN Learning Process
A trained CNN is a system capable of determining the exact
class of a given input data. The training is achieved through
an update of the layer’s parameters (filters, weights, and
biases) based on the error between the CNN predicted class
and the class label. The CNN learning process is an iterative
process based on the feedforward propagation and back-
propagation operations.
5.1. Feedforward Propagation. For the CNN model, the
feedforward equations can be derived from (1)–(5) and (6).
The Softmax activation [40, 41] function is applied in the
final layer to generate the predicted value of the class of the
input image I. For a multiclass model, the Softmax is
expressed as follows:
yi �
exp hi􏼁
􏽐c
j exp hj
􏼐 􏼑
, (7)
where c denotes the number of classes, yi is the ith
coor-
dinate of the output vector y, and the artificial neural output
hi � 􏽐n
j�1 hiwij.
5.2. Backpropagation. To update the CNN parameters and
perform the learning process, a backpropagation optimi-
zation algorithm is developed to minimize a selected cost
function E. In this analysis, the cross-entropy (CE) cost
function [40] is used.
E � − 􏽘
p
j�1
�
yi log yi􏼁 + 1 + �
yi􏼁log 1 − yi􏼁􏼁, (8)
here, �
yi is the desired output (data label).
The most used optimization algorithm to solve classi-
fication problems is the gradient descent (GD). Various
optimizers for the GD algorithm such as momentum,
AdaGrad, RMSprop, Adam, AdaMax, and Nadam were used
to improve the CNN performance.
5.2.1. Gradient Descent [40, 42]. GD is the simplest form of
optimization gradient descent algorithms. It is easy to im-
plement and gives significant classification accuracy. The
general update equation of the CNN parameters using the
GD algorithm is
φ(t + 1) � φ(t) − α
zE
zφ(t)
, (9)
where φ represents the update of the filters F, the weights W,
and the biases B. (zE/zφ) is the gradient with respect to the
parameter φ. α is the model learning rate. A too-large value
of α may lead to the divergence of the GD algorithm and may
cause the oscillation of the model performance. A too-small
α stops the learning process.
5.2.2. Gradient Descent with Momentum [43]. The mo-
mentum hyperparameter m defines the velocity by which the
learning rate α must be increased when the model ap-
proaches to the minimal of the cost function E. The update
equations using the momentum GD algorithm are expressed
as follows:
Input Layer
(Image)
(32 × 32)
(28 × 1)
Output
(Softmax)
(3 × 3)
16
CNN 1st
Layer
(3 × 3)
CNN 2nd
Layer
16
(3 × 3)
CNN 3rd
Layer
32
(3 × 3)
CNN 4th
Layer
32
ELU ELU
ELU
ELU
Dropout
Dropout
ELU ELU
Flatten
(2 × 2)
max-
pooling
Fully Connected
1000
Fully Connected
1000
(2 × 2)
max-
pooling
Figure 3: Proposed CNN architecture.
0 0 0 0 0 0 0 0
0
0
0
0
0
0
0
0
0
0
0
F
C
0
0 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1 -2
-2
-2
-2
-2
-2
1
2
2
2
2
2
1
1
1
2
2
0
0
0
0 0
0 0
0 0
0 0
0 0
0 0 0
0
0
0
0
I
0
0
0
0
0
0
0
0 0 0 0 0 0
0
1
1 1 1 1
1
1
1
1
1
1
1
1
1 1 1 1 1 1
0 0 0
0
0
0
Figure 4: Example of 2D convolution operation [13, 31].

v(t) � m v(t − 1) +
zE
zφ
,
φ(t + 1) � φ(t) − α v(t),
(10)
where v(t) is the moment gained at tth
iteration.
5.2.3. AdaGrad [44]. In this algorithm, the learning rate is a
function of the gradient (zE/zφ). It is defined as follows:
α(t) �
β
��
�
G(t) + ε
􏽰 , (11)
where
β �
α(0)
�
ε
√ ,
G(t) � 􏽘
t
i�1
zE
zφ(t)
􏼠 􏼡
2
,
(12)
where ϵ is a small smoothing value used to avoid the division
by 0 and G(t) is the sum of the squares of the gradients
(zE/zφ(t)).
With a small magnitude of (zE/zφ), the value of α is
increasing. If (zE/zφ) is very large, the value of α is a
constant. AdaGrad optimization algorithm changes the
learning rate for each parameter at a given time t with
considering the previous gradient update. The parameter
update equation using AdaGrad is expressed as follows:
φ(t + 1) � φ(t) − α(t)
zE
zφ(t)
. (13)
5.2.4. AdaDelta [45]. The issue of AdaGrad is that with
much iteration the learning rate becomes very small which
leads to a slow convergence. To fix this problem, AdaDelta
algorithm proposed to take an exponentially decaying av-
erage as a solution, where
E G2
(t)
􏽨 􏽩 � cE G2
(t − 1)
􏽨 􏽩 +(1 − c)G2
(t),
Δθ(t) �
1
��
�
E G2
(t)
􏽨 􏽩 + ε
􏽱 G(t),
φ(t + 1) � φ(t) − αΔθ(t),
(14)
where E[G2
(t)] is the decaying average over past squared
gradients and c is a set usually around 0.9.
5.2.5. RMSprop [45, 46]. In reality, RMSprop is identical to
AdaDelta’s initial update vector, which we derived above:
E G2
(t)
􏽨 􏽩 � 0.9 E G2
(t − 1)
􏽨 􏽩 + 0.1G2
(t),
φ(t + 1) � φ(t) − αΔθ(t).
(15)
5.2.6. ADAM [17, 45, 46]. This gradient descent optimizer
algorithm computes the learning rate α based on two vectors:
r(t) � β1r(t − 1) + 1 − β1􏼁
zE
zφ(t)
,
v(t) � β2v(t − 1) + 1 − β2􏼁
zE
zφ(t)
􏼠 􏼡
2
,
(16)
where r(t) and v(t) are the 1st
and the 2nd
order moments
vectors. β1 and β2 are the decay rates. r(t − 1) and v(t − 1)
represent the mean and the variance of the previous
gradient.
When r(t) and v(t) are very small, a large step size is
needed for parameters update. To avoid this issue, a bias
correction value is added to r(t) and v(t).
􏽢
r(t) �
r(t)
1 − βt
1
􏼐 􏼑
,
􏽢
v(t) �
v(t)
1 − βt
2
􏼐 􏼑
,
(17)
where βt
1 is β1 power t and βt
2 is β2 power t.
The Adam update equation is expressed as follows:
φ(t + 1) � φ(t) − α
􏽢
r(t)
��
�
􏽢
v(t) + ε
􏽰 . (18)
-2
-2 -2
-2
-2
-1
-1
-1
-1
-1
-1
-1
-1 0 0 0
1
1
1 0 0
C
0 0 0 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1 1
1
1
1
1
1
0 0
0 0
0 0
0 0
2
2
2
2
2
2
-1
Ca
Relu
Figure 5: Convolution map after applying ReLU activation
function [13, 25, 31].
Max (1, 2, 0, 1)=2
Average (5, 2, 0, 1) =2
1
1
1
1
0
0
0
0
2 2
2
2
1
5
5
7
3
2
2
2
5
5
3
7
Figure 6: Average-pooling, and max-pooling with filter (2 × 2)
and stride 2 [13, 31].

5.2.7. AdaMax [45, 47]. The factor v(t) in the Adam al-
gorithm adjusts the gradient inversely proportionate to the
ℓ2 norm of previous gradients (via the v(t − 1)) and current
gradient t (zE/zφ(t)) :
v(t) � β2v(t − 1) + 1 − β2􏼁
zE
zφ(t)
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
2
. (19)
The generalization of this update to the ℓp norm is as
follows:
v(t) � β
p
2 v(t − 1) + 1 − β
p
2
􏼐 􏼑
zE
zφ(t)
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
p
. (20)
To avoid being numerically unstable, ℓ1 and ℓ2 norms
are most common in practice. However, in general ℓ∞ also
shows stable behavior. As a result, the authors propose
AdaMax and demonstrate that v(t) with ℓ∞ converges to
the more stable value. Here,
u(t) � β∞
2 v(t − 1) + 1 − β∞
2 􏼁
zE
zφ(t)
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
∞
� max β2. v(t − 1),
zE
zφ(t)
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼌
􏼠 􏼡,
φ(t + 1) � φ(t) − α
􏽢
r(t)
u(t)
.
(21)
5.2.8. Nadam [43]. It is a combination of Adam and NAG,
where the parameters update equation using NAG is defined
as follows:
v(t) � c v(t − 1) + α
zE
zφ(t)
− c v(t − 1)
􏼠 􏼡,
φ(t + 1) � φ(t) − αv(t).
(22)
The update equation using Nadam is expressed as
follows:
r � β1􏽢
r(t) +
1 − β1􏼁
1 − β(t)1
zE
zφ(t)
􏼠 􏼡,
φ(t + 1) � φ(t) −
α
��
�
􏽢
v(t) + ε
􏽰 r.
(23)
6. Data Augmentation Techniques
Deep convolutional neural networks are heavily reliant on
big data to achieve excellent performance and avoid the
overfitting problem.
To solve the problem of insufficient data for Arabic
handwritten characters, we present some basic data aug-
mentation techniques that enhance the size and quality of
training datasets.
The image augmentation approaches used in this study
include geometric transformations, feature space augmen-
tation, noise injection, and mixing images.
Data augmentation based on geometric transformations
and feature space augmentation [17, 48] is often related to
the application of rotation, flipping, shifting, and zooming.
6.1. Rotation. The input data is rotated right or left on an
axis between 1° and 359°. The rotation degree parameter has
a significant impact on the safety of the dataset. For example,
on digit identification tasks like MNIST, slight rotations like
1 to 20 or −1 to −20 could be useful, but when the rotation
degree increases, properly the CNN network cannot accu-
rately distinguish between some digits.
6.2. Flipping. The input image is flipped horizontally or
vertically. This augmentation is one of the simplest to im-
plement and has proven useful on some datasets such as
ImageNet and CIFAR-10.
6.3. Shifting. The input image is shifting right, left, up, or
down. This transformation is a highly effective adjustment to
prevent positional bias. Figure 7 shows an example of
shifting data augmentation technique using Arabic alphabet
characters.
6.4. Zooming. The input image is zooming, either by adding
some pixels around the image or by applying random zooms
to the image. The amount of zooming has an influence on the
quality of the image; for example, if we apply a lot of
zooming, we can lose some image pixels.
6.5. Noise Injection. As it could be seen on Arabic hand-
written characters, natural noises are presented in images.
Noises make recognition more difficult and for this reason,
noises are reduced by image preprocessing techniques. The
cos of noise reduction is to perform a high classification, but
it causes the alteration of the character shape. The main
datasets in this research topic are considered with denoising
images. The question which we answer here is how the
method could be robust to any noise.
Adding noise [48, 49] to a convolution neural network
during training helps the model learn more robust features,
resulting in better performance and faster learning. We can
add several types of noise when recognizing images, such as
the following.
(i) Gaussian noise: injecting a matrix of random values
drawn from a Gaussian distribution
(ii) Salt-and-pepper noise: changing randomly a certain
amount of the pixels to completely white or com-
pletely black
(iii) Speckle noise: only adding black pixels “pepper” or
white pixels “salt”
Adding noise to the input data is the most commonly
used approach, but during training, we can add random
noise to other parts of the CNN model. Some examples
include the following:

(i) Adding noise to the outputs of each layer
(ii) Adding noise to the gradients to update the model
parameters
(iii) Adding noise to the target variables
6.6.MixingImage’sDatabases. In this study, we augment the
training dataset by mixing two different Arabic handwritten
characters datasets, AHCD and Hijja, respectively. AHCD is
a clean database, but Hijja is a dataset with very low-reso-
lution images. It comprises many distorted alphabets
images.
Then, we evaluate the influence of different mentioned
data augmentation techniques (geometric transformations,
feature space augmentation, and noise injection) on the
recognition performance of the new mixing dataset.
7. Experimental Results and Discussion
7.1. Datasets. In this study, two datasets of Arabic hand-
written characters were used: Arabic handwritten characters
dataset “AHCD” and Hijja dataset.
AHCD [6] comprises 16.800 handwritten characters of
size (32 × 32 ×1) pixels. It was written by 60 participants
between the ages of 19 and 40 years and most of the par-
ticipants are right handed. Each participant wrote the Arabic
alphabet from “alef” to “yeh” 10 times. The dataset has 28
classes. It is divided into a training set of 13.440 characters
and a testing set of 3.360 characters.
Hijja dataset [13] consists of 4.434 Arabic characters of
size (32 × 32 ×1) pixels. It was written by 591 school children
ranging in age between 7 to 12 years. Collecting data from
children is a very hard task. Malformed characters are
characteristic of children’s handwriting; therefore the
dataset comprises repeated letters, missing letters, and many
distorted or unclear characters. The dataset has 29 classes. It
is divided into a training set of 37.933 characters and a
testing set of 9.501 characters (80% for training and 20% for
test).
Figure 8 shows a sample of AHCD and Hijja Arabic
handwritten letters datasets.
7.2. Experimental Environment and Performance Evaluation.
In this study the implementation and the evaluation of the
CNN model are done out in Keras deep learning environ-
ment with TensorFlow backend on Google Colab using GPU
accelerator.
We evaluate the performance of our proposed model via
the following measures:
Accuracy (A) is a measure for how many correct
predictions your model made for the complete test
dataset:
A �
TP + TN
TP + TN + FN + FP
. (24)
Recall (R) is the fraction of images that are correctly
classified over the total number of images that belong to
class:
R �
TP
TP + FN
. (25)
Precision (P) is the fraction of images that are correctly
classified over the total number of images classified:
P �
TP
TP + FP
. (26)
F1 measure is a combination of Recall and Precision
measures:
F1 � 2 ∗
P ∗ R
P + R
. (27)
Here, TP � true positive (is the total number of images
that can be correctly labeled as belonging to a class x), FP �
false positive (represents the total number of images that
have been incorrectly labeled as belonging to a class x),
FN � false negative (represents the total number of images
that have been incorrectly labeled as not belonging to a class
x), TN � true negative (represents the total number of images
that have been correctly labeled as not belonging to a class x).
Also we draw the area under the ROC curve (AUC),
where we have the following.
An ROC curve (receiver operating characteristic curve)
is a graph showing the performance of all classification
thresholds. This curve plots two parameters:
(i) True-positive rate
(ii) False-positive rate
AUC stands for “area under the ROC curve.” That is,
AUC measures the entire two-dimensional area underneath
the entire ROC curve from (0.0) to (1.1).
7.3. Tuning of CNN Hyperparameters. The objective is to
choose the best model that fits the AHCD and Hijja datasets
well. Many try-and-error trials in the network configuration
tuning mechanism were performed.
The best performance was achieved when the CNN
model was constructed of four convolution layers followed
by three fully connected hidden layers. The model starts with
0
0 20
10
20
30
0
0 20
10
20
30
0
0 20
10
20
30
0
0 20
10
20
30
0
Down Shifted Up Shifted
0 20
10
20
30
Shifted Left
Shifted Right
Original
Figure 7: Data augmentation through shifting a single data in four
directions by two pixels.

two convolution layers with 16 filters of size (3 × 3), then the
remaining 2 convolution layers are with 32 filters of size
(3 × 3), and each two convolution layers are followed by
max-pooling layers with (2 × 2) kernel dimension. Finally,
three fully connected layers (dense layers) with Softmax
activation function to perform prediction. ELU, a nonlinear
activation function, was used to remove negative values by
converting them into 0.001. The values of weights and bias
are updated by a backward propagation process to minimize
the loss function.
To reduce the overfitting problem a dropout of 0.6 rate is
added to a model between the dense layers and applies to
outputs of the prior layer that are fed to the subsequent layer.
The optimized parameters used to improve the CNN per-
formance were as follows: Optimizer algorithm is Adam, the
loss function is the cross-entropy, learning rate � 0.001,
batch size � 16, and epochs � 40.
We compare our model to CNN-for-AHCD over both
the Hijja dataset and the AHCD dataset. The code for CNN-
for-AHCD is available online [31], which allows comparison
of its performance over various datasets.
On the Hijja dataset, which has 29 classes, our model
achieved an average overall test set accuracy of 88.46%,
precision of 87.98%, recall of 88.46%, and an F1 score of
88.47%, while CNN-for-AHCD achieved an average overall
test set accuracy of 80%, precision of 80.79%, recall of
80.47%, and an F1 score of 80.4%.
On the AHCD dataset, which has 28 classes, our model
achieved an average overall test set accuracy of 96.66%,
precision of 96.75%, recall of 96.67%, and an F1 score of
96.67%, while CNN-for-AHCD achieved an average overall
test set accuracy of 93.84%, precision of 93.99%, recall of
93.84%, and an F1 score of 93.84%.
The detailed metrics are reported per character in
Table 3.
We note that our model outperforms CNN-for-AHCD
by a large margin on all metrics.
Figure 9 shows the testing result AUC of AHCD and
Hijja dataset.
7.4.OptimizerAlgorithms. The objective is to choose the best
optimizers algorithms that fit the AHCD and Hijja best
performance. In this context, we tested the influence of the
following algorithms on the classification of handwritten
Arabic characters:
(i) Adam
(ii) SGD
(iii) RMSprop
(iv) AdaGrad
(v) Nadam
(vi) Momentum
(vii) AdaMax
By using Nadam optimization algorithm, on the Hijja
dataset, our model achieved an average overall test set ac-
curacy of 88.57%, precision of 87.86%, recall of 87.98%, and
an F1 score of 87.95%.
On the AHCD dataset, our model achieved an average
overall test set accuracy of 96.73%, precision of 96.80%,
recall of 96.73%, and an F1 score of 96.72%.
The detailed results of different optimizations algorithms
are mentioned in Table 4.
7.5. Results of Data Augmentation Techniques. Generally, the
neural network performance is improved through the good
tuning of the model hyperparameters. Such improvement in
the CNN accuracy is linked to the availability of training
dataset. However, the networks are heavily reliant on big
data to avoid overfitting problem and perform well.
Data augmentation is the solution to the problem of
limited data. The image augmentation techniques used and
discussed in this study include geometric transformations
and feature space augmentation (rotation, shifting, flipping,
and zooming), noise injection, and mixing images from two
different datasets.
For the geometric transformations and feature space
augmentation, we try to well choose the percentage of
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
(a)
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
0
0
20
20
(b)
Figure 8: Arabic handwritten letters dataset. (a) AHCD dataset. (b) Hijja dataset.

Table 3: Experimental results on the Hijja and AHCD datasets via CNN-for-AHCD model.
Characters
Hijja dataset AHCD dataset
CNN-for-AHCD Our model CNN-for-AHCD Our model
P R F1 P R F1 P R F1 P R F1
1. Alif 0.93 0.97 0.95 0.97 0.98 0.98 0.96 0.99 0.98 0.96 1.00 0.98
2. Baa 0.82 0.91 0.86 0.93 0.96 0.94 0.97 0.97 0.97 0.99 0.97 0.98
3. Tea 0.66 0.88 0.75 0.84 0.92 0.88 0.87 0.95 0.91 0.96 0.95 0.95
4. Thea 0.76 0.81 0.78 0.94 0.93 0.93 0.95 0.88 0.92 0.97 0.95 0.96
5. Jam 0.79 0.85 0.82 0.80 0.90 0.84 0.95 0.96 0.95 0.99 0.98 0.99
6. Haa 0.83 0.60 0.70 0.85 0.89 0.87 0.93 0.93 0.93 0.99 0.96 0.97
7. Khaa 0.76 0.77 0.77 0.82 0.73 0.77 0.94 0.93 0.93 0.99 0.96 0.97
8. Daal 0.65 0.69 0.67 0.79 0.68 0.73 0.91 0.94 0.93 0.95 0.88 0.92
9. Thaal 0.70 0.68 0.69 0.83 0.92 0.87 0.96 0.91 0.93 0.88 0.95 0.92
10. Raa 0.86 0.87 0.87 0.81 0.91 0.86 0.89 0.98 0.94 0.94 0.99 0.96
11. Zaay 0.87 0.89 0.88 0.96 0.90 0.93 0.94 0.88 0.91 0.96 0.92 0.94
12. Seen 0.84 0.92 0.88 0.95 0.92 0.94 0.95 0.91 0.93 1.00 0.97 0.99
13. Sheen 0.86 0.82 0.84 0.90 0.88 0.89 0.92 0.98 0.95 0.99 1.00 1.00
14. Sad 0.75 0.81 0.78 0.90 0.86 0.88 0.84 0.96 0.90 0.96 0.97 0.97
15. Dhad 0.80 0.76 0.78 0.91 0.93 0.92 1.00 0.89 0.94 0.97 0.95 0.96
16. Tah 0.90 0.83 0.87 0.96 0.89 0.92 0.96 0.94 0.95 0.94 0.98 0.96
17. Dha 0.83 0.87 0.85 0.84 0.82 0.83 0.97 0.94 0.95 0.97 0.95 0.96
18. Ain 0.74 0.70 0.71 0.86 0.86 0.86 0.95 0.90 0.92 0.98 0.98 0.98
19. Ghen 0.83 0.71 0.77 0.75 0.86 0.80 0.89 0.97 0.93 0.98 0.99 0.99
20. Fa 0.77 0.65 0.71 0.92 0.87 0.89 0.92 0.84 0.88 0.94 0.98 0.96
21. Qaf 0.81 0.80 0.81 0.87 0.89 0.88 0.87 0.91 0.89 0.99 0.95 0.97
22. Kaf 0.86 0.78 0.82 0.93 0.89 0.91 0.98 0.96 0.97 0.98 0.93 0.96
23. Lam 0.90 0.87 0.88 0.91 0.89 0.90 0.98 0.97 0.98 0.99 0.98 0.99
24. Meem 0.83 0.85 0.84 0.82 0.80 0.81 0.98 0.98 0.98 0.98 0.98 0.98
25. Noon 0.70 0.77 0.73 0.88 0.83 0.86 0.92 0.92 0.92 0.86 0.97 0.91
26. Ha 0.81 0.76 0.78 0.89 0.94 0.91 0.97 0.96 0.96 0.97 0.96 0.97
27. Waw 0.930 0.82 0.87 0.94 0.92 0.93 0.96 0.94 0.95 0.97 0.94 0.95
28. Yaa 0.82 0.81 0.82 0.87 0.82 0.84 0.97 0.97 0.97 0.99 0.98 0.99
29. Hamza 0.74 0.73 0.74 na na na na na na
Acc. (train) 0.88 0.98 0.91 1.00
Acc. (test) 0.80 0.88 0.94 0.96
Macro avg 0.81 0.80 0.80 0.88 0.88 0.88 0.94 0.94 0.94 0.97 0.96 0.96
Weighted avg 0.81 0.80 0.80 0.89 0.88 0.88 0.94 0.94 0.94 0.97 0.96 0.96
0.0
0.2
0.4
True
Positive
Rate
0.6
0.8
1.0
Receiver Operating Characteristic Curve
0.0 0.2 0.4 0.6
False Positive Rate
0.8
AUC = 0.999
1.0
(a)
0.0
0.2
0.4
True
Positive
Rate
0.6
0.8
1.0
Receiver Operating Characteristic Curve
0.0 0.2 0.4 0.6
False Positive Rate
0.8 1.0
AUC = 0.996
(b)
Figure 9: AUC curve of Arabic handwritten characters dataset. (a) AUC of AHCD testing dataset. (b) AUC of Hijja testing dataset.

Table
4:
Experimental
results
on
the
Hijja
and
AHCD
through
diﬀerent
optimizers
algorithms.
Algorithm
Accuracy
(%)
Precision
(%)
Recall
(%)
F1
score
(%)
AHCD
Hijja
AHCD
Hijja
AHCD
Hijja
AHCD
Hijja
Training
Testing
Training
Testing
Training
Testing
Training
Testing
Training
Testing
Training
Testing
Training
Testing
Training
Testing
Adam
99.76
96.66
97.60
88.46
99.76
96.75
97.40
87.98
99.76
96.67
97.60
88.46
99.76
96.67
97.60
88.47
SGD
99.44
95.68
94.16
87.85
99.45
95.84
94.28
87.93
99.45
95.68
94.16
87.85
99.45
95.70
94.17
87.88
RMSprop
99.72
96.22
96.55
87.98
99.73
96.30
94.28
87.86
99.72
96.22
94.16
87.98
99.72
96.23
94.17
87.95
AdaGrad
88.62
83.54
63.90
60.37
88.73
83.78
87.86
64.95
88.62
83.54
87.98
63.91
88.60
83.52
87.95
63.77
Nadam
99.77
96.73
97.76
88.57
99.77
96.80
97.57
87.86
99.77
96.73
97.76
87.98
99.77
96.72
97.76
87.95
Momentum
99.16
95.56
94.16
87.76
99.18
95.72
94.11
87.73
99.17
95.57
94.17
87.77
99.17
95.58
94.15
87.74
AdaMax
99.89
96.19
97.19
87.57
99.90
96.25
97.01
87.86
99.90
96.19
97.19
87.98
99.90
96.19
97.20
87.95

rotation, shifting, flipping, and zooming for the model
attending a good performance. For example, if we rotate the
Latin handwritten number database (MNIST) by 180°, the
network will not be able to accurately distinguish between
the handwritten digits “6” and “9”. Likewise, on the AHCD
and Hijja datasets, if rotating or flipping techniques are
used the network will be unable to distinguish between
some handwritten Arabic characters. For example, as
shown in Figure 10, with a rotation of 180°, the character
Daal isolated (‫)د‬ will be the same as the character Noon
isolated (‫)ن‬.
The detailed results of rotation, shifting, flipping, and
zooming data augmentation techniques are mentioned in
Table 5.
As shown in Table 5 and Figure 11, by using rotation and
shifting augmentation approaches, our model achieved a
testing accuracy of 98.48% and 91.24% on AHCD dataset
and Hijja dataset, respectively. We achieved this accuracy
through rotating the input image by 10° and shifting it just by
one pixel.
Adding noise is a technique used to augment the training
input data. Also in most of the cases, this is bound to increase
the robustness of our network.
In this work we used the three types of noise to augment
our data:
(i) Gaussian noise
(ii) Salt-and-pepper noise
0
30
20
10
0
Daal_isolated
20 0
30
20
10
0
Noon_isolated
20
(a)
0
30
20
10
0
Lam beginning
20 0
30
20
10
0
Alif end
20
(b)
Figure 10: Example of Arabic handwritten characters that cannot be properly distinguished when applying the rotation and flipping data
augmentation techniques. (a) Confusion caused by rotation. (b) Confusion caused by flipping.
Table 5: Experimental results of data augmentation techniques on Hijja and AHCD datasets.
Dataset
Data augmentation techniques with using Nadam algorithm
Rotation Shifting Rotation and shifting Flipping Zooming
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
AHCD 99.82 97.64 99.56 98.09 99.41 98.48 99.69 97.29 99.85 98.00
Hijja 97.18 89.86 92.72 90.28 91.73 91.24 95.39 88.94 99.16 88.66
0.5
0 5 10 15 20
Epoch
25 30 35 40
0.6
0.7
0.8
Accuracy
0.9
1.0
train accuracy
test accuracy
train accuracy augmented
test accuracy augmented
(a)
0.5
0 5 10 15 20
Epoch
25 30 35 40
0.6
0.7
0.8
Accuracy
0.9
0.4
train accuracy
test accuracy
train accuracy augmented
test accuracy augmented
(b)
Figure 11: Rotation and shifting data augmentation results of Arabic handwritten characters datasets. (a) Rotation and shifting results of
AHCD dataset. (b) Rotation and shifting results of Hijja dataset.

(iii) Speckle noise
The detailed results of different types of noise injection
are mentioned in Table 6. As shown by adding different types
of noise, the model accuracy is improved, which demon-
strate the robustness of our proposed architecture. We
achieved good results when adding noise to the outputs of
each layer.
The proposed idea in this study is to augment the
number of training databases by mixing the two datasets
AHCD and Hijja, and then we apply the previously men-
tioned data augmentation methods on the new mixed
dataset. Our purpose to use malformed handwritten char-
acters as it proposes the Hijja dataset is to improve the
accuracy of our method with noised data.
The detailed results of data augmentation techniques on
the mixed database are mentioned in Table 7. As shown, the
model performance depends on the rate of using Arabic
handwriting “Hijja” database. The children had trouble
following the reference paper, which results in very low-
resolution images comprising many unclear characters.
Therefore mixing the datasets would certainly reduce
performance.
8. Conclusions and Possible Future
Research Directions
In this paper, we proposed a convolution neural network
(CNN) to recognize Arabic handwritten characters dataset.
We have trained the model on two Arabic datasets AHCD
and Hijja. By the good tuning of the network
hyperparameters, we achieved an accuracy of 96.73% and
88.57% on AHCD and Hijja.
To improve the model performance, we have imple-
mented different optimization algorithms. For both data-
bases, we achieved an excellent performance by using
Nadam optimizer.
To solve the problem of insufficient Arabic handwritten
datasets, we have applied different data augmentation
techniques. The augmentation approaches are based on
geometric transformation, feature space augmentation,
noise injection, and mixing of datasets.
By using rotation and shifting techniques, we achieved a
good accuracy equal to 98.48% and 91.24% on AHCD and
Hijja.
To improve the robustness of the CNN model and in-
crease the number of training datasets, we added three types
of noise (Gaussian noise, Salt-and-pepper, and Speckle
noise).
Also in this work we first augmented the database by
mixing two Arabic handwritten characters datasets; then we
tested the results of the previously mentioned data aug-
mentation techniques on the new mixed dataset, where the
first database “AHCD” comprises clear images with a very
good resolution, but the second database “Hijja” has many
distorted characters. Experimentally show that the geo-
metric transformations (rotation, shifting, and flipping),
feature space augmentation, and noise injection always
improve the network performance, but the rate of using the
unclean database “Hijja” harms the model accuracy.
An interesting future direction is the cleaning and
processing of Hijja dataset to eliminate the problem of low-
Table 7: Experimental results on mixing of Hijja and AHCD through different data augmentation techniques.
Mixed dataset
Data augmentation techniques with using Nadam algorithm
Rotation Shifting Rotation and shifting Flipping Gaussian noise
Training Testing
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
Training
accuracy
Testing
accuracy
80(%):
AHCD
20(%):
Hijja
20(%):
AHCD
10(%):
Hijja
99.62 97.42 99.36 98.02 99.38 98.32 99.77 97.08 99.15 96.74
80(%):
Hijja
20(%):
AHCD
20(%):
Hijja
10(%):
AHCD
97.36 88.47 93.66 89.07 94.91 90.54 95.21 88.21 96.68 88.49
80(%):
AHCD
80(%):
Hijja
20(%):
AHCD
20(%):
Hijja
97.27 74.53 97.16 75.13 98.13 78.13 98.16 74.22 96.98 74.02
Table 6: Experimental results on the Hijja and AHCD through noise injection.
Type of noise injection with using Nadam algorithm
Gaussian noise Salt-and-pepper noise Speckle noise
Training accuracy Testing accuracy Training accuracy Testing accuracy Training accuracy Testing accuracy
AHCD 99.17 96.78 99.16 97.15 99.14 96.82
Hijja 97.18 89.85 92.96 90.10 91.32 89.73

resolution and unclear images and then the implementation
of the proposed CNN network and data augmentation
techniques on the new mixed and cleaned database.
In addition, we are interested in evaluating the result of
other augmentation approaches, like adversarial training,
neural style transfer, and generative adversarial networks on
the recognition of Arabic handwritten characters dataset.
We plan to incorporate our work into an application for
children that teaches Arabic spelling.
Abbreviations
AHCR: Arabic handwritten characters recognition
DL: Deep learning
CNNs: Convolution neural networks
AHCD: Arabic handwritten character dataset
SVM: Support vector machine
ADBase: Arabic digits database
HACDB: Handwritten Arabic characters database
OIHACDB: Offline handwritten Arabic character database
CDCGAN: Conditional deep convolutional generative
adversarial network
Tanh: Hyperbolic tangent
ReLU: Rectified linear unit
CE: Cross-entropy
GD: Gradient descent
NAG: Nesterov accelerated gradient
TP: True positive
FP: False positive
FN: False Negative
TN: True negative
AUC: Area under curve
ROC: Receiver operating curve
ELU: Exponential linear unit
Symbols
I : Image
m: Width and height of the image
c: Number of channels
F: Filter
n: Filter size
⊗: Convolution operation
C: Convolution map
a: Size of convolution map
Sc: Stride
p: Padding
f: Nonlinear activation function
Ca: Convolution map after applying f
K: Kernel
Sp: Number of patches
pool: Pooling operation
P: Pooling map
Y0
: Concatenation vector
Pc
i : Output of the convolution layer
Pc−1
i−1 : Convoluted image
Yi− 1
: Input of the i th fully connected hidden layer
Yi
: Output of the i th fully connected hidden layer
Hi
: Weight sum vector
Bi
: Bias
E: Cost function
�
yi: Desired output
ϕ: Update of the filter F
(zE/zφ): Gradient
α: Model learning
m: Momentum
v(t): Moment gained at the i th iteration
ε: Smoothing value
G(t): Sum of the squares of the gradient
E[G(t)2
]: Decaying overage
r(t): Moments vector
β: Decay rate
r(t − 1): Mean of the previous gradient
v(t − 1): Variance of the previous gradient.
Data Availability
Previously reported AHCD data were used to support this
study and are available at https://www.kaggle.com/mloey1/
ahcd1. These prior studies (and datasets) are cited at relevant
places within the text as [43].
Conflicts of Interest
The authors declare that there are no conflicts of interest
regarding the publication of this study.
References
[1] Y. M. Alginahi, “Arabic character segmentation: a survey,”
International Journal on Document Analysis and Recognition,
vol. 16, no. 2, pp. 105–126, 2013.
[2] M. T. Parvez and S. A. Mahmoud, “Offline arabic handwritten
text recognition: a survey,” ACM Computing Surveys, vol. 45,
no. 2, pp. 1–35, 2013.
[3] D. K. Sahu and C. V. Jawahar, “Unsupervised feature learning
for optical character recognition,” in Proceedings of the 2015
13th International Conference on Document Analysis and
Recognition (ICDAR), pp. 1041–1045, Tunis, Tunisia, August
2015.
[4] J. Sueiras, V. Ruiz, A. Sanchez, and J. F. Velez, “Offline
continuous handwriting recognition using sequence to se-
quence neural networks,” Neurocomputing, vol. 289,
pp. 119–128, 2018.
[5] J. Bai, Z. Chen, B. Feng, and B. Xu, “Image character rec-
ognition using deep convolutional neural network learned
from different languages,” in Proceedings of the 2014 IEEE
International Conference on Image Processing (ICIP),
pp. 2560–2564, Paris, France, October 2014.
[6] A. El-Sawy, M. Loey, and E. Hazem, “Arabic handwritten
characters recognition using convolutional neural network,”
WSEAS Transactions on Computer Research, vol. 5, pp. 11–19,
2017.
[7] J. Xiao, Z. Xuehong, H. Chuangxia, Y. Xiaoguang,
W. Fenghua, and M. Zhong, “A new approach for stock price
analysis and prediction based on SSA and SVM,” Interna-
tional Journal of Information Technology Decision Making,
vol. 18, pp. 287–310, 2019.
[8] N. Wagaa and H. Kallel, “Vector-based back propagation
algorithm of supervised convolution neural network,” in
Proceedings of the International Conference on Control,

Automation and Diagnosis (ICCAD), Paris, France, October
2020.
[9] Y. Liang, J. Wang, S. Zhou, Y. Gong, and N. Zheng, “In-
corporating image priors with deep convolutional neural
networks for image super-resolution,” Neurocomputing,
vol. 194, pp. 340–347, 2016.
[10] H. M. Najadat, A. A. Alshboul, and A. F. Alabed, “Arabic
handwritten characters recognition using convolutional
neural network,” in Proceedings of the 10th International
Conference on Information and Communication Systems,
Irbid, Jordan, June 2019.
[11] H. Kaur and S. Rani, “Handwritten gurumukhi character
recognition using convolution neural network,” International
Journal of Computational Intelligence Research, vol. 13,
pp. 933–943, 2017.
[12] P. Ahamed, S. Kundu, T. Khan, V. Bhateja, R. Sarkar, and
A. Faruk Mollah, “Handwritten arabic numerals recognition
using convolutional neural network,” Journal of Ambient In-
telligence and Humanized Computing, vol. 11, pp. 5445–5457,
2020.
[13] N. Altwaijry and I. Al-Turaiki, “Arabic handwriting recog-
nition system using convolutional neural network,” Neural
Computing Applications, vol. 33, 2020.
[14] K. V. Greeshma and K. Sreekumar, “Hyperparameter opti-
mization and regularization on fashion-MNIST classifica-
tion,” International Journal of Recent Technology and
Engineering (IJRTE), vol. 8, pp. 2277–3878, 2019.
[15] O. Y. Bakhteev and V. V. Strijov, “Comprehensive analysis of
gradient-based hyperparameter optimization algorithms,”
Annals of Operations Research, vol. 289, 2020.
[16] S. Derya, “A comparison of optimization algorithms for deep
learning,” International Journal of Pattern Recognition and
Artificial Intelligence, vol. 34, pp. 51–65, 2020.
[17] A. H. Garcı́a and P. König, “Further advantages of data
augmentation on convolutional neural networks,” in Pro-
ceedings of the International Conference on Artificial Neural
Networks, pp. 95–103, Springer, Cham, Switzerland, Sep-
tember 2018.
[18] A. A. Hidayata, K. Purwandaria, T. W. Cenggoroa, and
B. Pardamean, “A convolutional neural network-based an-
cient sundanese character classifier with data augmentation,”
in Proceedings of the 5th International Conference on Com-
puter Science and Computational Intelligence, pp. 195–201,
Indonesia, 2021.
[19] K. Younis, “Arabic handwritten characters recognition based
on deep convolutional neural networks,” Jordan Journal
Computers and Information Technology (JJCIT), vol. 3, 2018.
[20] G. Latif, J. Alghazo, L. Alzubaidi, M. M. Naseer, and
Y. Alghazo, “Deep convolutional neural network for recog-
nition of unified multi-language handwritten numerals,” in
Proceedings of the 2018 IEEE 2nd International Workshop on
Arabic and Derived Script Analysis and Recognition (ASAR),
pp. 90–95, London, UK, March 2018.
[21] N. Alrobah and S. Albahli, “A hybrid deep model for rec-
ognizing arabic handwritten characters,” IEEE Acces, vol. 9,
pp. 87058–87069, 2021.
[22] M. A. Mudhsh and R. Almodfer, “Arabic handwritten al-
phanumeric character recognition using very deep neural
network,” MDPI, Information, vol. 8, 2017.
[23] C. Boufenar, A. Kerboua, and M. Batouche, “Investigation on
deep learning for off-line handwritten arabic character rec-
ognition,” Cognitive Systems Research, vol. 50, 2017.
[24] I. B. Mustapha, S. Hasan, H. Nabus, and S. M. Shamsuddin,
“Conditional deep convolutional generative adversarial
networks for isolated handwritten arabic character genera-
tion,” Arabian Journal for Science and Engineering, 2021.
[25] A. Yuan, G. Bai, L. Jiao, and Y. Liu, “Offline handwritten
English character recognition based on convolutional neural
network,” in Proceedings of the 2012 10th IAPR International
Workshop on Document Analysis Systems (DAS), pp. 125–129,
IEEE, Gold Coast, Australia, March 2012.
[26] M. Mhapsekar, P. Mhapsekar, A. Mhatre, and V. Sawant,
“Implementation of residual network (ResNet) for devanagari
handwritten character recognition,” Algorithms for Intelligent
Systems, pp. 137–148, 2020.
[27] N. K. Pius and A. Johny, “Malayalam handwritten character
recognition system using convolutional neural network,”
International Journal of Applied Engineering Research, vol. 15,
pp. 918–920, 2020.
[28] R. S. Hussien, A. A. Elkhidir, and M. G. Elnourani, “Optical
character recognition of arabic handwritten characters using
neural network,” in Proceedings of the 2015 International
Conference on Computing, Control, Networking, Electronics
and Embedded Systems Engineering (ICCNEEE), pp. 456–461,
IEEE, Khartoum, Sudan, September 2015.
[29] A. ElAdel, R. Ejbali, M. Zaied, and C. B. Amar, “Dyadic
multiresolution analysis-based deep learning for arabic
handwritten character classification,” in Proceedings of the
2015 IEEE 27th International Conference on Tools with Ar-
tificial Intelligence (ICTAI), pp. 807–812, IEEE, Vietri sul
Mare, Italy, November 2015.
[30] M. Elleuch, N. Tagougui, and M. Kherallah, “Arabic handwritten
characters recognition using deep belief neural networks,” in
Proceedings of the 2015 12th International Multi-Conference on
Systems, Signals & Devices (SSD), pp.1–5, IEEE, Mahdia, Tunisia,
March 2015.
[31] A. Elsawy, M. Loey, and H. M El-Bakry, “Arabic handwritten
characters recognition using convolutional neural network,”
WSEAS Transactions on Computer Research, vol. 5, pp. 11–19,
2017.
[32] K. M. M. Yaagoup and M. E. M. Mus, “Online Arabic
handwriting characters recognition using deep learning,”
International Journal of Advanced Research in Computer and
Communication Engineering, vol. 9, 2020.
[33] M. Shams, A. Elsonbaty, and W. ElSawy, “Arabic handwritten
character recognition based on convolution neural networks
and support vector machine,” International Journal of Ad-
vanced Computer Science and Applications, vol. 11, 2020.
[34] S. Narayan, “The generalized sigmoid activation function:
competitive supervised learning,” Information Sciences,
vol. 99, no. 1-2, pp. 69–82, 1997.
[35] B. L. Kalman and C. Kwasny, “Why tanh: choosing a sig-
moidal function,” in Proceedings of the JCNN International
Joint Conference on Neural Networks, IEEE, Baltimore, MD,
USA, June 1992.
[36] B. I. Dlimi and H. Kallel, “Robust Neural Control for Robotic
Manipulators,” International journal of Enhanced Research in
Science, Technology and Engineering, IJERSTE, vol. 5, no. 2,
pp. 198–205, 2016.
[37] P. Sibi, S. Allwyn Jones, and P. Siddarth, “Analysis of different
activation functions using back propagation neural net-
works,” Journal of Theoretical and Applied Information
Technology, vol. 47, 2013.
[38] G. Lin and W. Shen, “Research on convolutional neural
network based on improved relu piecewise activation func-
tion,” Procedia Computer Science, vol. 131, pp. 977–984, 2018.
[39] M. Kubat, “Artificial neural networks,” An Introduction to
Machine Learning, pp. 91–111, 2015.

[40] L. Bottou, “Large-scale machine learning with stochastic
gradient descent,” Proceedings of COMPSTAT’2010, pp. 177–
186, 2010.
[41] D. Wang, L. Huang, and L. Tang, “Dissipativity and synchro-
nization of generalized BAM neural networks with multivariate
discontinuous activations,” IEEE Transactions on Neural Net-
work Learning System, vol. 29, pp. 3815–3827, 2017.
[42] R. Sutton, “Two problems with back propagation and other
steepest descent learning procedures for networks,” in Pro-
ceedings of the Eighth Annual Conference of the Cognitive
Science Society, pp. 823–832, Chicago, IL, USA, December
1986.
[43] I. Sutskever, J. Martens, G. Dahl, and G. Hinton, “On the
importance of initialization and momentum in deep learn-
ing,” in Proceedings of the International Conference on Ma-
chine Learning, pp. 1139–1147, Atlanta, GA, USA, June 2013.
[44] J. Duchi, E. Hazan, and Y. Singer, “Adaptive sub gradient
methods for online learning and stochastic optimization,”
Journal of Machine Learning Research, vol. 12, no. Jul,
pp. 2121–2159, 2011.
[45] S. Derya, “A comparison of optimization algorithms for deep
learning,” International Journal of Pattern Recognition and
Artiﬁcial Intelligence, 2020.
[46] G. Hinton, N. Srivastava, and K. Swersky, rmsprop: Divide the
Gradient by a Running Average of its Recent Magnitude, 2012.
[47] D. P. Kingma and L. J. Ba, “Adam: a method for stochastic
optimization,” in Proceedings of the International Conference
on Learning Representations (ICLR), San Diego, CA, USA,
May 2015.
[48] S. Connor and M. K. Taghi, “A survey on image data aug-
mentation for deep learning,” Journal of Big Data, vol. 6, 2019.
[49] P. Panda and K. Roy, “Implicit advers arial data augmentation
and robustness with noise-based learning,” Neural Networks,
vol. 141, pp. 1139–1147, 2021.

Improved Arabic Alphabet Characters Classification Using.pdf

More Related Content

Similar to Improved Arabic Alphabet Characters Classification Using.pdf (20)

Recently uploaded (20)

Improved Arabic Alphabet Characters Classification Using.pdf