Adi Alhudhaif

Bio: Adi Alhudhaif is an academic researcher from Salman bin Abdulaziz University. The author has contributed to research in topics: Computer science & Artificial intelligence. The author has an hindex of 3, co-authored 28 publications receiving 34 citations. Previous affiliations of Adi Alhudhaif include George Washington University.

Robust automated Parkinson disease detection based on voice signals with transfer learning

Abstract: Parkinson's disease (PD) is a progressive-neurodegenerative disorder that affects more than 6 million people around the world. However, conventional techniques for PD detection are often hand-crafted, in which special expertise is needed. In this study, considering the importance of rapid diagnosis of the disease, it was aimed to develop deep convolutional neural networks (CNN) for automated PD identification based on biomarkers-derived voice signals. The developed CNN methods consisted of two main stages, including data pre-processing and fine-tunning-based transfer learning steps. To train and evaluate the performance of the developed model, datasets were collected from the mPower Voice database. SqueezeNet1_1, ResNet101, and DenseNet161 architectures were retrained and evaluated to determine which architecture can classify frequency-time information most accurately. The performance results revealed that the proposed model could successfully identify the PD with an accuracy of 89.75%, sensitivity of 91.50%, and precision of 88.40% for DenseNet-161 architecture identified as the most suitable fine-tuning architecture. The results revealed that the proposed model based on transfer learning with a fine-tuning approach provides an acceptable detection of PD with an accuracy of 89.75%. The outcomes of the study confirmed that by integrating the developed model into smart electronic devices, it will be able to develop alternative pre-diagnosis methods and will assist the physicians for PD detection during the in-clinic assessment. The success of the proposed model would imply an enhancement in the life quality of patients and a cost reduction for the national health system.

Determination of COVID-19 Pneumonia based on Generalized Convolutional Neural Network Model from Chest X-Ray Images

Abstract: X-ray units have become one of the most advantageous candidates for triaging the new Coronavirus disease COVID-19 infected patients thanks to its relatively low radiation dose, ease of access, practical, reduced prices, and quick imaging process. This research intended to develop a reliable convolutional-neural-network (CNN) model for the classification of COVID-19 from chest X-ray views. Moreover, it is aimed to prevent bias issues due to the database. Transfer learning-based CNN model was developed by using a sum of 1,218 chest X-ray images (CXIs) consisting of 368 COVID-19 pneumonia and 850 other pneumonia cases by pre-trained architectures, including DenseNet-201, ResNet-18, and SqueezeNet. The chest X-ray images were acquired from publicly available databases, and each individual image was carefully selected to prevent any bias problem. A stratified 5-fold cross-validation approach was utilized with a ratio of 90% for training and 10% for the testing (unseen folds), in which 20% of training data was used as a validation set to prevent overfitting problems. The binary classification performances of the proposed CNN models were evaluated by the testing data. The activation mapping approach was implemented to improve the causality and visuality of the radiograph. The outcomes demonstrated that the proposed CNN model built on DenseNet-201 architecture outperformed amongst the others with the highest accuracy, precision, recall, and F1-scores of 94.96%, 89.74%, 94.59%, and 92.11%, respectively. The results indicated that the reliable diagnosis of COVID-19 pneumonia from CXIs based on the CNN model opens the door to accelerate triage, save critical time, and prioritize resources besides assisting the radiologists.

Classification of imbalanced hyperspectral images using SMOTE-based deep learning methods

Abstract: Hyperspectral imaging (HSI) is one of the most advanced methods of digital imaging This technique differs from RGB images with its wide range of the electromagnetic spectrum Imbalanced data sets are frequently encountered in machine learning As a result, the classifier performance may be poor To avoid this problem, the data set must be balanced The main motivation in this study is to reveal the difference and effects on the classifier performance between the original imbalanced dataset and the data set modified by balancing methods In the proposed method, hyperspectral image classification study carried out on Xuzhou Hyspex dataset includes nine-classes including bareland-1, bareland-2, crops-1, crops-2, lake, coals, cement, trees, house-roofs of elements, by using the convolutional neural networks (CNN) and dataset balancing methods comprising the Smote, Adasyn, K-Means, and Cluster This dataset has been taken from IEEE-Dataport Machine Learning Repository To classify the hyperspectral image, the convolutional neural networks having different multiclass classification approaches like One-vs-All, One-vs-One Dataset was splitted in two different ways: %50–%50 Hold-out and 5-Fold Cross-validation In order to evaluate the performance of the proposed models, the confusion matrix, classification accuracy, precision, recall, and F-Measure have been used Without the dataset balancing, the obtained classification accuracies are 9363%, 9233%, 8836% for %50–%50 train-test split, and 9446%, 94%, 9224% for 5-Fold cross-validation using multi-class classification, One-vs-All, and One-vs-One respectively After Smote balancing, the obtained classification accuracies are 9641%, 956%, 9253% for %50–%50 train-test split and 9649%, 9564%, 9338% for 5-Fold cross-validation using multi-class classification, One-vs-All and One-vs-One respectively After Adasyn balancing, the obtained classification accuracies are 9586%, 9362%, 8705% for %50–%50 train-test split and 9638%, 9509%, 9155% for 5-Fold cross-validation using multi-class classification, One-vs-All and One-vs-One respectively After K-Means balancing, the obtained classification accuracies are 9523%, 9336%, 906% for %50–%50 train-test split and 9574%, 9472%, 9194% for 5-Fold cross-validation using multi-class classification, One-vs-All and One-vs-One respectively After Cluster balancing, the obtained classification accuracies are 9483%, 941%, 9007% for %50–%50 train-test split and 9628%, 9588%, 925% for 5-Fold cross-validation using multi-class classification, One-vs-All and One-vs-One respectively The obtained results have shown that the best model is Smote Balanced 5-CV multiclass classification

Prediction of electric vehicle charging duration time using ensemble machine learning algorithm and Shapley additive explanations

Abstract: Electric vehicles (EVs) are the most important components of smart transportation systems. Limited driving range, prolonged charging times, and inadequate charging infrastructure are the key barriers to EV adoption. To address the problem of prolonged charging time, the simple approach of developing a new charging station to enhance the charging capacity may not work due to the limitation of physical space and strain on power grids. Prediction of precise EV charging time can assist the drivers in effective planning of their trips to alleviate range anxiety during trips. Therefore, this study employed four different ensemble machine learning (EML) algorithms: random forest, extreme gradient boosting (XGBoost), categorical boosting, and light gradient boosting machine, for predicting EVs' charging time. The prediction experiments were based on 2 years of real‐world charging event data from 500 EVs in Japan's private and commercial vehicles. The study emphasized predicting charging time for different charging modes, that is, normal and fast charging operations. The results indicate that EML models performed well under various scenarios, with the XGBoost model having the highest accuracy. Moreover, we also employ the newly developed Shapley additive explanation (SHAP) approach to tackle the non‐interpretability issues of the ML algorithm by interpreting the XGBoost model outputs. The obtained SHAP value plots demonstrated the nonlinear relationship between explanatory variables and EV charging time.

Sensitive sandwich-type electrochemical SARS-CoV­2 nucleocapsid protein immunosensor.

Abstract: A sensitive and fast sandwich-type electrochemical SARS-CoV‑2 (COVID-19) nucleocapsid protein immunosensor was prepared based on bismuth tungstate/bismuth sulfide composite (Bi2WO6/Bi2S3) as electrode platform and graphitic carbon nitride sheet decorated with gold nanoparticles (Au NPs) and tungsten trioxide sphere composite (g-C3N4/Au/WO3) as signal amplification. The electrostatic interactions between capture antibody and Bi2WO6/Bi2S3 led to immobilization of the capture nucleocapsid antibody. The detection antibody was then conjugated to g-C3N4/Au/WO3 via the affinity of amino-gold. After physicochemically characterization via transmission electron microscopy (TEM), scanning electron microscopy (SEM), x-ray diffraction (XRD), and x-ray photoelectron spectroscopy (XPS), cyclic voltammetry (CV), differential pulse voltammetry (DPV), and electrochemical impedance spectroscopy (EIS) analysis were implemented to evaluate the electrochemical performance of the prepared immunosensor. The detection of SARS-CoV-2 nucleocapsid protein (SARS-CoV-2 NP) in a small saliva sample (100.0 µL) took just 30 min and yielded a detection limit (LOD) of 3.00 fg mL−1, making it an effective tool for point-of-care COVID-19 testing.

Real-time RGB image encryption for IoT applications using enhanced sequences from chaotic maps

Abstract: Four chaotic maps are used herein as case study to design an embedded cryptosystem based on a pseudo-random number generator (PRNG). The randomness of the sequences is enhanced by applying the m o d 1023 function and verified by analyzing bifurcation diagrams, the maximum Lyapunov exponent, and performing NIST SP 800-22 and TestU01 statistical tests. The PRNG is applied in a simple algorithm for real-time RGB images encryption on a machine-to-machine (M2M) scheme, using message queuing telemetry transport (MQTT) protocol over WiFi network and through Internet. The cryptanalysis confirms that the proposed image encryption scheme is robust to resist most of the existing attacks, such as statistical histograms, entropy, key-space, correlation of adjacent pixels, and differential attacks. The implementation of the proposed cryptosystem is done using enhanced sequences from the Logistic 1D map, and it reaches a throughput of up to 47.44 Mbit/s using a personal computer with a 2.9 GHz clock, and 10.53 Mbit/s using a Raspberry Pi 4. As a result, our proposed embedded cryptosystem is suitable to increase the security in the transmission of RGB images in real-time through WiFi networks and Internet.

A Cross-Layer Approach MAC/NET with Updated-GA (MNUG-CLA)-Based Routing Protocol for VANET Network

Abstract: Nowadays, technology is developed rapidly in communication technology. Several new technologies have been introduced due to the evolution of wireless communication and this provided the way to communicate among vehicles, using a Vehicular Ad-Hoc Network (VANETs). Routing in VANETs becomes most challenging because of the huge mobility and dynamical topology changes, which lead to reduced efficiency in the network. The core idea of this network is to increase the efficiency during the process of the communication. The most suited routing protocol for VANETs is Geographic routing, for the reason that it provides higher scalability and low overheads. The major challenges in VANETs are the selection of best neighbor in dynamically changing VANET topology. Furthermore, to provide better QoS needful actions are essential. In this paper, we introduced a new MAC/NET with Updated Genetic Algorithm—A Cross Layer Approach, (MNUG-CLA) based on a MAC layer and network layer to overcome the drawbacks of the network. In the network layer, a new neighbor discovery protocol is developed to select the best next hop for the dynamically varying network. In the MAC layer, in order to improve the quality, multi-channel MAC model is introduced for instantaneous transmission from various service channels. For overall optimal path selection, we used an updated GA algorithm. The performance was demonstrated through the use of an extensive simulation environment, NS-2. The simulation results prove that this protocol provides better results, in terms of energy efficiency, energy consumption and successive packet transmission, when compared with the earlier approaches.