IRAQI JOURNAL OF COMPUTERS, COMMUNICATIONS, CONTROL AND SYSTEMS ENGINEERING

since the global pandemic of COVID-19 has spread out, the use of Artificial
Intelligence to analyze Chest X-Ray (CXR) image for COVID-19 diagnosis and patient
treatment is becoming more important. This research hypothesized that using COVID19
radiographic changes in the X-Ray images. Artificial Intelligence (AI) systems may extract
certain graphical elements regarding COVID-19 and offer a clinical diagnosis ahead of
pathogenic test; therefore, saving vital time for disease prevention. Employing 2614 CXR
radiographs from Kaggle data collection of verified COVID-19 cases and healthy persons,
a new Convolutional Neural Network (CNN) model that is inspired by the Xception
architecture was presented for the diagnosis of coronavirus pneumonia infected patients.
The suggested technique reached an average validation accuracy of 0.99, precision of 0.95,
recall of 0.92, and F1-score of 0. 95. Finally, such findings revealed that the Deep Learning
(DL) technique has the potential to decrease frontline radiologists' stress, enhance early
diagnosis, treatment, and isolation; therefore, aid in epidemic control.

Hackers can conduct more destructive cyber-attacks thanks to the rapid spread
of Internet of Things (IoT) devices, posing significant security risks for users. Through a
malicious process, the attacker intended to exhaust the capital of the target IoT network.
Researchers and company owners are concerned about the reliability of IoT networks,
which is taken into account because it has a significant impact on the delivery of facilities
provided by IoT systems and the security of user groups. The intrusion prevention system
ensures that the network is protected by detecting malicious activity. In this paper, the focus
is on predicting attacks and distinguishing between normal network use and network
exploitation for intrusion and network attack and we will use Swarm Intelligence (SI) which
is one of the types of artificial intelligence (AI) that we harness to choose features to
determine the task of them and specifically we will use an algorithm Meerkat Clan (MCA)
for this purpose, as this research suggested a modified IDS in machine learning (ML) based
IoT environments to identify features and these features will be input into Random Forest
algorithm. The IoTID20 dataset is used where nominal traits are removed, so the final
dataset contains 79 traits. The data set contains three categories: the label that identifies
whether it is a natural use or exploitation, the category that characterizes the type of
exploitation, and the subcategory that describes that exploitation more accurately. The
number of trees in a random forest (RF) classifier for binary, class, and subclass will be
determined by the experiment. The trained classifier is then tested and the approach
achieves 100% accuracy for binary target prediction, 96.5% for category and accuracy
ranges of 83.7% for sub-category target prediction. The proposed system is evaluated and
compared with previous systems and its performance is shown through the use of confusion
matrix and others.

Traditional identification techniques for electronic payments, such
as the Personal Identification Number (PIN), are becoming outdated and
unsafe, while mobile payments are becoming more popular and widely used. It
presents a risk to issuers since there is no reliable consumer verification method
available, and the lack of safe and reliable e-payment systems is one of the key
issues restricting progress. As a result, efforts have been made to develop and
maintain a unified payment system that is well-organized, efficient, dependable,
and secure. This system avoids the need for physical cash while also still
satisfying all payment and identification requirements, a safe and trustworthy
method is required for the country's successful adoption of an e-payment
system. This article focuses on the future of online payment and the security
problems through using effective biometric authentication technologies to
provide a trustworthy authentication method for an e-payment system.

The paper presents an observer-based estimation of sensor fault for
control systems affected by friction force. In such systems, the non-linearity of
friction force leads to deteriorating sensor fault estimation capability of the
observer. Hence, the challenge is to design an observer capable of attaining
robust sensor fault estimation while avoiding the effects of friction. To
overcome the highlighted challenge, an Unknown Input Observer (UIO) is
designed to decouple the effects of friction as well as to estimate the state and
sensor fault.The benefit of proposing UIO is to guarantee robust sensor fault
estimation despite the highly non-linear disturbance in the form of friction. The
gains of the UIO are computed through a single–step linear matrix inequality.
Finally, an inverted pendulum simulation is presented to demonstrate the novel
approach's performance effectiveness.

Content-Based Image Retrieval (CBIR) is an actual application in
computer vision, which retrieves similar images from a database. Deep
Learning (DL) is essential in many applications, including image retrieval
applications. However, encryption techniques are used to protect data privacy
because these data are vulnerable to being viewed by unauthorized parties
while being transmitted over unsecured channels.
This paper includes two parts for images retrieval. In the first part, features of
all images of a Canadian Institute for Advanced Research CIFAR-10 dataset
were extracted and stored on the Server-side. In the second part, the
Brakerski/Fan-Vercauteren (BFV) homomorphic encryption scheme method for
encrypting an image sent by the client-side. First, their decryption and image
features are extracted depending on the trainer model when they arrive on the
server-side. Then an extracted features are compared with stored features using
the Cosine Distance method, and then the server encrypts the retrieved images
and sends them to the client-side. Deep-learning results on plain images were
97% for classification and 96.7% for retriever images. At the same time,
The National Institute of Standards and Technology (NIST ) test was used to
check the security of BFV when applied to CIFAR-10 dataset.

Network traffic has risen in recent years to the point that it is
obviously and astonishingly in 2020, with the increase predicted to double in
the following days. Up to 23 Teraa bit every month is an incredible amount. The
Active Queue Management (AQM) algorithm is one of the most significant study
areas in network congestion control; nevertheless, new self-learning network
management algorithms are needed on nodes to cope with the huge quantity of
traffic and minimize queuing latency, used reinforcement learning for automatic
adaptive parameter with the AQM algorithm for effective network management,
and present a novel AQM algorithm that focuses on deep reinforcement
learning to deal with latency and the trade-off between queuing delay and
throughput; choose Deep Q-Network (DQN) as the foundation for our scheme
and equate it with Random Early Detection (RED) Results based on Network
simulation (NS3) simulation suggest that the DQN algorithm has good and
better results were obtained from RED, where the difference reached a drop
rate of 2%, and this percentage is considered good, in addition to the
percentage of throughput and the packet transfer rate of 3% is better in the
proposed algorithm.

Road traffic accidents are one of the leading causes of mortality globally.
Reducing the number of traffic-related incidents has become a serious socio-economic and
public health problem, given the ever-increasing number of cars on the road. As a result,
this paper proposes an intelligent vehicle prediction communication mechanism that alerts
drivers to any autos that may be overtaking or bypassing the targeted vehicle. The primary
goal of this paper is to leverage modern Internet of Things (IoT) and wireless sensor
technologies to predict any potential accident that may occur as a result of car accidents.
This paper proposes the Collision Prediction of a Moving Vehicle (CPMV) system. The
information acquired by CPMV will alert the driver to divert the vehicle in a reasonable
amount of time before any harm occurs. It redirects the inbound object that emitted the
Ultrasound signal which was received by the vehicle, to a safe location. The proposed
system predicts collision between vehicles through Wi-Fi and Bluetooth, using a set of
sensors with a precision of 360 degrees and a distance of collision prediction of one meter
and at a speed of 200-300 revolutions per minute. The python programming language was
utilized to code the programs that control the vehicle during the implementation of this
project. The Raspberry Pi 4 is utilized as the controller to examine the vehicle’s spatial
data. The test results showed that using this application to deal with an approaching object
can be a successful strategy in the three proposed scenarios at different angles and
directions.

Many academics have concentrated on applying machine learning to
retrieve information from databases to enable researchers to perform better. A
difficult issue in prediction models is the selection of practical strategies that
yield satisfactory forecast accuracy. Traditional software testing techniques
have been extended to testing machine learning systems; however, they are
insufficient for the latter because of the diversity of problems that machine
learning systems create. Hence, the proposed methodologies were used to
predict flight prices. A variety of artificial intelligence algorithms are used to
attain the required, such as Bayesian modeling techniques such as Stochastic
Gradient Descent (SGD), Adaptive boosting (ADA), Decision Trees (DT), K-
nearest neighbor (KNN), and Logistic Regression (LR), have been used to
identify the parameters that allow for effective price estimation. These
approaches were tested on a data set of an extensive Indian airline network.
When it came to estimating flight prices, the results demonstrate that the
Decision tree method is the best conceivable Algorithm for predicting the price
of a flight in our particular situation with 89% accuracy. The SGD method had
the lowest accuracy, which was 38 %, while the accuracies of the KNN, NB,
ADA, and LR algorithms were 69 %, 45 %, and 43 %, respectively. This study's
presented methodologies will allow airline firms to predict flight prices more
accurately, enhance air travel, and eliminate delay dispersion.

Feature space high dimensionality is a well-known problem in text
classification and web mining domains, it is caused mainly by the large number
of vocabularies contained within web documents. Several methods were applied
to select the most useful and important features over the years; however, the
performance of such methods is still improvable from different aspects such as
the computational cost and accuracy. This research presents an enhanced
cosine similarity-based hybridization of two efficient feature selection methods
for higher classification performance. The reduced feature sets are generated
using the Random Projection (RP) and the Principal Component Analysis
(PCA) methods, individually, then hybridized based on the cosine similarity
values between features’ vectors. The performance of the proposed method in
terms of accuracy and F-measure was tested on a dataset of web pages based
on several term weighting schemes. As compared to relevant methods, results of
the proposed method show significantly higher accuracy and f-measure
performance based on less feature set size.

The main reason for the emergence of a deepfake (deep learning and fake) term
is the evolution in artificial intelligence techniques, especially deep learning. Deep
learning algorithms, which auto-solve problems when giving large sets of data, are used to
swap faces in digital media to create fake media with a realistic appearance. To increase
the accuracy of distinguishing a real video from fake one, a new model has been developed
based on deep learning and noise residuals. By using Steganalysis Rich Model (SRM)
filters, we can gather a low-level noise map that is used as input to a light Convolution
neural network (CNN) to classify a real face from fake one. The results of our work show
that the training accuracy of the CNN model can be significantly enhanced by using noise
residuals instead of RGB pixels. Compared to alternative methods, the advantages of our
method include higher detection accuracy, lowest training time, a fewer number of layers
and parameters.

In recent years, especially with COVID-19, video conference
applications have become very important. Millions of peoples around the world
have become to communicate with each other through using video conference
applications. The most critical factor in the performance success of video
conference applications is the user's perception of the quality of the experience.
In this work, an Extreme Learning Machine (ELM) model was proposed for
predicting video quality of experience. The proposed system extracts several
features from videos that have a significant impact on the quality of the
experience. The model performance was validated with unseen data.
Spearman’s Rank Correlation Coefficient (SRCC), Root Mean Square Error
(RMSE), Pearson’s Linear Correlation Coefficient (PLCC) metrics have been
used to measure the accuracy of the model and correlation. The results
demonstrate that the proposed model had better performance than models used
by the previous researchers that were used for predicting video QoE in terms of
precision, correlation, and running time.

Artificial Neural Networks (ANNs) are utilized to solve a variety of problems in
many domains. In this type of network, training and selecting parameters that define
networks architecture play an important role in enhancing the accuracy of the network's
output; Therefore, Prior to training, those parameters must be optimized. Grey Wolf
Optimizer (GWO) has been considered one of the efficient developed approaches in the
Swarm Intelligence area that is used to solve real-world optimization problems. However,
GWO still faces a problem of the slump in local optimums in some places due to insufficient
diversity. This paper proposes a novel algorithm Levy Flight- Chaotic Chen mapping on
Wolf Pack Algorithm in Neural Network. It efficiently exploits the search regions to detect
driving sleepiness and balance the exploration and exploitation operators, which are
considered implied features of any stochastic search algorithm. Due to the lack of dataset
availability, a dataset of 15 participants has been collected from scratch to evaluate the
proposed algorithm's performance. The results show that the proposed algorithm achieves
an accuracy of 99.3%.

Thomas and Patrik are working on a stream cipher called SNOW
3G. In 2006, it was chosen as the centerpiece of a new set of confidentiality and
integrity algorithms for the Universal Mobile Telecommunications System
(UMTS). In 2008, Böhm published an article named "Statistical Evaluation of
Stream Cipher SNOW 3G." He put the randomness of the SNOW 3G key stream
generator to the test. As a randomness test tool, Böhm uses the NIST test
statistics package, which consists of three kinds of tests: lengthy key stream
data, short key stream data, and initialization vector data. Only the short key
stream set of data failed eight random chance test results out of three kinds of
tests, according to the report's findings. The SNOW 3G suggestions, he claims,
fail due to a flaw with in cipher's initialization. In this paper, we use the PLL
algorithm to modify the SNOW 3G algorithm for key initialization and
generation keystream. We employ the same itself Böhm employed. The modify
SNOW 3G algorithm exceed whole of the statistic exam in a experiment. The
findings show that PLL has an effect on algorithm randomness.

The Internet is a massive network that connects millions of users from all
over the world and the data transmitted via it needs great protection, especially since that
are in the age of big data. To solve part of this problem, IPsec was utilized, which is a set
of protocols necessary to offer security to units of the Internet in general and the IP layer
in particular. It is mostly based on major exchange protocols. The most frequent
mechanism for transferring key materials and establishing security linkages between two
entities is Internet Key Exchange (IKE). In the present work, it is proposed to use a public
key that works together with Diffie-Hellman cryptography and the main advantages of a
single-stage contribution (as opposed to the two-stage in standard IKE) it is better in
terms of improved transfer and time (more time for the corresponding negotiation) to
make the proposed IKE more secure with Simple account constraints