Hybrid Cyber-Security Model for Attacks Detection Based on Deep and Machine Learning

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2022, Номер 18(11), С. 17 - 30

Опубликована: Авг. 31, 2022

Nowadays, numerous attacks can be considered high risks in terms of the security Wireless Sensor Networks (WSN). As a result, different applications are introduced to manage data and information exchange related sides save transmission data. Recently, most classified as cyber ones. These interest system halting destroying rather than stealing In this paper, cyber-attacks detection is proposed based on an intelligent hybrid model that uses deep machine learning technologies. The improves cyber-attack speed. addition, feature reduction using methods (PCA SVD) select features adopted classes attacks. This positively affect deep-learning complexity. obtained results demonstrate superiority model-based comparison traditional ones reaching accuracy 99.98%, 100%, 100% for precision, recall, F1-measure respectively, reducing time 23s datasets Message Queuing Telemetry Transport-Dataset (MQTT-DS) Dataset (WSN-DS).

Язык: Английский

---

An Image Feature Extraction to Generate a Key for Encryption in Cyber Security Medical Environments

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(01), С. 93 - 106

Опубликована: Янв. 17, 2023

Cyber security is a term utilized for describing collection of technologies, procedures, and practices that try protecting an online environment user or organization. For medical images among most important delicate data kinds in computer systems, the reasons require all patient data, including images, be encrypted before being transferred over networks by healthcare companies. This paper presents new direction encryption method research encrypting image based on domain feature extracted to generate key process. The process started applying edges detection. After dividing bits edge into (3×3) windows, diffusions are applied create used image. Four randomness tests passed through NIST ensure whether generated accepted as true. reversible state decryption retrieve original will gained can any cyber field such comparative experiments prove proposed algorithm improves efficiency has good performance, higher information entropy 7.42 well lower correlation coefficient 0.653.

Язык: Английский

---

Adaptive Hiding Algorithm Based on Mapping Database

International Journal of Interactive Mobile Technologies (iJIM), Год журнала: 2023, Номер 17(01), С. 96 - 107

Опубликована: Янв. 10, 2023

Information hiding one of the important field security which provide secure level for information. Achieving multi levels system often researchers used cryptography side by with steganography. Utilizing message digest algorithm to play role crypto is extracted from secret created database. Message (MD5) two times as one-way function data integrity. The implemented evaluated based on peak signal noise ratio (PSNR) metric and best value reaches 62.46. proposed works in adaptive behavior due different use images well selected point could be generate hash code well. up sufficient through using both steganography cryptography.

Язык: Английский

---

The Implementation of Online Learning in Conventional Higher Education Institutions During the Spread of COVID-19: A Comparative Study

International Journal of Emerging Technologies in Learning (iJET), Год журнала: 2023, Номер 18(01), С. 68 - 99

Опубликована: Янв. 10, 2023

The purpose of this study is to investigate and explore the degree success implementation online learning in conventional higher education institutions instead face-to-face during spread Covid-19 Pandemic 2019/2020 academic year, via exploring undergraduate students' perceptions application system at Ajman University UAE, Griffith Australia. In study, descriptive approach was used. A questionnaire consisting 40 items designed distributed 630 students from 675 University, who were randomly selected different faculties two universities year COVID-19 pandemic. results revealed that a moderate satisfaction with University's readiness, training, technical support for university's teaching process pandemic, female finding them more than male students. Disciplines computer skills also showed an impact on such satisfaction, Pharmacy & Health Science College Architecture, Art, Design discipline those excellent both Universities. addition, positive attitudes towards use

Язык: Английский

---

Comparison YOLOv5 Family for Human Crowd Detection

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(04), С. 94 - 108

Опубликована: Апрель 3, 2023

Recent years have seen widespread application of crowd counting and detection technology in areas as varied urban preventing crime, station statistics, people flow studies. However, getting accurate placements improving audience performance dense scenes still has challenges, it pays to devote a lot effort it. In this paper, models are proposed based on the YOLOv5 algorithm, four (YOLOv5l, YOLOv5m, YOLOv5s, YOLOv5x) were built for purpose comparing increasing accuracy identification each model contains certain characteristics such Filter sizes. Each was trained human dataset (indoor outdoor) results showing which reaches higher detecting people. Through study practical experiments conducted model, found that best is YOLOv5x, YOLOv5l, where humans reached more than 96%, while YOLOv5s 92%, YOLOv5m lowest accuracy, 91%.

Язык: Английский

---

Intrusion Detection in Wireless Body Area Network using Attentive with Graphical Bidirectional Long-Short Term Memory

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(06), С. 31 - 46

Опубликована: Май 16, 2023

Recently developed low-power networked systems, wireless communications, and sensors have all contributed to the rise of Wireless Sensor Networks (WSNs) as a potentially useful tool in medical field. Securing Body Area (WBANs) is essential for their widespread use healthcare environments because data they send frequently includes private confidential patient health information. The study's goal create system detecting intrusions WBAN. To best identify attacks such we present novel “Attention-based Bi-directional Long Short-Term Memory with Graph Construction” (ABL-GC) here. suggested approach ensures that intrusion detection uses only features detect given attack, reducing processing complexity.

Язык: Английский

---

An Efficient System for Diagnosis of Human Blindness Using Image-Processing and Machine-Learning Methods

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(10), С. 82 - 98

Опубликована: Авг. 1, 2023

The two main causes of blindness are diabetes and glaucoma. Routine diagnosis is based on the conventional robust mass-screening method. However, despite being cost-effective, this method has some problems as a human eye-disease detection because there many types eye disease that similar or result in no visual changes image. These issues make it highly difficult to recognize control it. Moreover, color macula spot can be very close affected variety diseases, which suggests indicate various possibilities, rather than one. This paper discusses shortcomings current blindness-screening monitoring systems presents feature-based approach using digital fundus images for purpose automated disorders, considering three conditions: healthy eye, diabetic retinopathy (DR), As such, develops computer-aided (CAD) blindness. proposed integrates Gabor filter features, statistical colored morphological local binary pattern then compares them with features drawn from standard dataset 1580 images. Several classification techniques were applied extracted-features neural network (NN), support vector machine (SVM), naïve bias (NB). SVM classifiers show most promising accuracy. They achieved 93.3% over other classifiers.

Язык: Английский

---

Investigation of Early-Stage Breast Cancer Detection using Quantum Neural Network

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(03), С. 61 - 81

Опубликована: Март 14, 2023

aided image diagnostics (CAD) have been used in many fields of diagnostic medicine. It relies heavily on classical computer vision and artificial intelligence. Quantum neural network (QNN) has introduced by researchers around the world presented recently research corporations such as Microsoft, Google, IBM. In this paper, investigation validity using QNN algorithm for machine-based breast cancer detection was performed. To validate learnability QNN, a series tests were performed alongside with convolutional (CCNN). is built Cirq library to perform assimilation quantum computation computers. Series investigations study characteristics CCNN under same computational conditions. The comparison real Mammogram data sets. showed success terms recognizing training. Our work shows better performance successfully training producing valid model smaller set compared CCNN.

Язык: Английский

---

Text to speech using Mel-Spectrogram with deep learning algorithms

Periodicals of Engineering and Natural Sciences (PEN), Год журнала: 2022, Номер 10(3), С. 380 - 380

Опубликована: Июнь 29, 2022

The purpose of text to speech (TTS), sometimes called synthesis, is synthesize a natural and intelligible for given text. A wide range applications uses TTS technologies in media, chatbots, entertainment, among other fields, making it hot topic the research community. Recently, progress achieved by artificial intelligence, especially deep learning neural networks, enables produce high-quality synthesized speech. However, despite success achieved, currently, available works suffer from need very long training inference time, which makes dominated big tech companies. This paper proposes model based on convolutional networks (CNN) gated recurrent units (GRU). proposed can work even low computational environments requires time. MOS 4.26, higher than performed state-of-the-art methods.

Язык: Английский

---

A robust facemask forgery detection system in video

Periodicals of Engineering and Natural Sciences (PEN), Год журнала: 2022, Номер 10(3), С. 212 - 212

Опубликована: Июнь 16, 2022

An in-depth fake video uses an Artificial Intelligent (AI), AI programming, and a Personal computer (PC) mix to create deep of the action. Deep-faking can also be used represent images sounds. We provide insights into our reviews in this document. We're showing dataset start. At point, we present subtleties reproductively exploratory settings evaluate discovered effects finally. It is no surprise find videos, which only monitor tiny section (e.g., target face appears quickly on video; hence time limited). remove system's fixed duration's persistent as each contributes preparation, approval, testing sections reflect this. The edge groups are isolated from successively (without outline skips). entire pipeline ready finished when approval stage ten years old. Convolutional Neural Network (CNN) was best most reliable classification systems. Fake videos typically use low-quality pictures mask faults or insist that general public regard camera defects unexplainable phenomena. 'This common trope with Unidentified Flying Object (UFO) videos: ghostly orbs lenses; snakes compression artifacts one's face. In study, have implemented sophisticated, knowledgeable method recognize false images. Our test results using various monitored shown reliably predict whether through simple co-evolutionary Long Short-Term Memory (LSTM) structure.

Язык: Английский

---