الباحثين

الكلية

عنوان البحث بلغة البحث

ملخص البحث

روان محمد يوسف غالى

     أ.د/هبة يوسف محمد

د/ أمانى أحمد مصطفى

د/ محمد شعبان حماده

كلية طب

Cardiac Profile Assessment of Acute Opioid Intoxicated Patients

روان أحمد محمد الداودى

أ.د/هبة يوسف محمد

د/ محمد شعبان حماده

د/ منى إبراهيم اليمانى

كلية طب

Poison Severity Score of Acute Pediatric Intoxication Among Cases Presenting to Pediatric Department In Damietta General Hospital

نيهال جمال شرف الدين

أ.د/إيناس إبراهيم الشيخ

أ.د/ نسرين سعد فراج

د/ نسرين فاروق محمد

كلية طب

Civic Engagement Effect on Adolescents’ Behaviors and Mental Health: A Comparative Cross-sectional Study

آية السيد فوزي علي رزق

أ.د/ محمد عيسى

أ.دمصطفى حراجى

آية السيد

العلوم

A Robust Ensemble Deep Learning Framework for Breast Cancer Image Retrieval

The increasing reliance on digital imaging in the medical field has led to significant challenges in storing, organizing, and retrieving large-scale image datasets. Content-Based Medical Image Retrieval (CBMIR) systems have emerged as effective tools to facilitate image access based on visual content. However, their performance is often constrained by the semantic gap between low-level image features and the clinical understanding required for diagnosis. To address this issue, this study introduces a hybrid CBMIR framework that integrates multiple deep learning models to enhance both image classification and retrieval. The proposed model utilized class-level predictions to retrieve semantically relevant images. We used the hybrid proposed model that include the integration of three deep learning CNN, RNN and XAI and we achieved high performance accuracy is 99.24% and loss is 0.1926.

لؤي فوزي محمد العتر

حازم الزينى

جنا مرعى

آلاء أمين

عبد الرحمن الشاذلى

عبد الرحمن الجبالى

مصطفى حراجى

محمد حسن

العلوم

From Deep LSTM-Based Gene Expression Modeling to microRNA Disease Association; hsa-miR-133b identified as a Potentially Functionally Conserved Tumor Suppressor Across Cancers

Cancer is one of the leading causes of death globally, making early and accurate diagnosis crucial for effective treatment. While deep learning models have shown promising results in cancer detection, most studies have focused on binary classification, leaving a gap in multi-class classification. Additionally, the role of microRNA (miRNA) molecules in cancer development and their potential use as biomarkers requires further exploration. This study aimed to develop a hybrid deep learning model for accurate cancer classification (both binary and multi-class) using gene expression data, and to identify tumor-suppressive miRNA molecules across multiple cancer types. A hybrid model called Deep Long Short-Term Memory (D-LSTM) was developed and trained on gene expression profile data extracted from the GEO database, specifically from the dataset GSE203024. This dataset included 50,675 genes profiled across 2,845 peripheral blood samples, comprising 14 types of cancer, colon polyps, and normal samples. The age of the subjects ranged from 18 to 97 years, with a sex distribution of 1,547 males and 1,013 females. Data preprocessing involved the selection of significant genes using the ANOVA-F test and Correlation-based Feature Selection (CFS) technique. Class imbalance was addressed using the synthetic minority oversampling technique (SMOTE). The performance of the model was evaluated using metrics such as accuracy, precision, recall, and F1-score. Subsequently, key genes were analyzed to identify associated miRNA molecules using the miRNet database, and regulatory networks were visualized using Cytoscape. The model achieved 98.38% accuracy in binary classification (cancer vs. non-cancer) and 99.88–100% accuracy in multiclass classification. The analysis revealed that hsa-miR-133b functions as a tumor suppressor in seven distinct cancer types—including breast, lung, pancreatic, liver, ovarian, prostate, and colorectal cancers—by targeting key genes such as CTNND1, CCNB1, ZMYM2, and SUZ12.

Our model demonstrates high accuracy in cancer classification, positioning it as a promising noninvasive diagnostic tool. The identification of hsa-miR-133b as a conserved tumor suppressor highlights its significance in cancer research. Further biological validation is required to confirm the potential clinical applicability of this method.

Keywords: Cancer; Deep learning; Gene expression; MicroRNA

ايمان فتحى على سند نوفل

خالد الوكيل

مختار سامى بحيرى

أحمد عبد الله

العلوم

Synthesis of novel nano composites based on alginate beads incorporated metal oxides\ Sulphide for industrial wastewater treatment

Sodium alginate is one of such biodegradable polymers, which has been extensively exploited for the preparation of nano-particles (NPs) for several wastewater treatment applications. Many biopolymers are used as adsorbents and encapsulation of nano-particle onto them can increase their efficiency. In this study, zinc oxide/ Sulphide, Arabian gum (AG) and titanium oxide incorporated on alginates beads have been synthesized and characterized both microscopically using scanning electron microscope (SEM) and spectroscopally using energy dispersive X- ray (EDX) and X-ray diffraction. Also, zero-point charge (PZC) was measured for the prepared adsorbents to identify its charge. These were then being used for the removal of some selected heavy metals and methylene blue dye (MB) from wastewater. The adsorption kinetics of the selected metal ions and dye onto the prepared beads were determined using each of Langmuir, Freundlich, first, second pseudo order and temkin models to show the homogenous properties of the synthesized particles.  For metal ions removal; Adsorption experiments targeted Cd(II), Pb(II), and Ni(II), and the results showed significant variations in adsorption capacities: Alg\ZnS\ZnO beads achieved 2.11, 1.79, and 1.94 mmol g⁻¹ for Cd(II), Pb(II), and Ni(II), respectively; Alg\ZnS\ZnO\Ti2O3 beads exhibited capacities of 3.50, 2.36, and 2.49 mmol g⁻¹, while Alg\ZnS\ZnO\GA beads recorded 1.47, 1.25, and 1.39 mmol g⁻¹ for the same ions at pH 6, 25°C. Kinetic and isotherm models were employed to elucidate the adsorption mechanisms. Kinetic studies confirmed pseudo- second-order behavior, indicating chemisorption as the primary mechanism, while isothermal studies demonstrated monolayer adsorption. Practical application tests with industrial effluents highlighted the beads’ high efficacy, removing up to 99.8% of heavy metals. Furthermore, reusability assessments showed minimal performance loss over five cycles, affirming the materials’ economic and environmental viability.  For MB dye adsorption; The observed removal efficiency of the nano composites were 98.5%, 73.5% and 58.3% for (Alg\ZnO\ZnS\AG), (Alg\ZnO\ZnS) and (Alg\ZnO\ZnS\Ti2O3) respectively with optimized pH of 1-15, contact time of 240 minute and agitation speed of 150 rpm and temperature of 25°. The adsorption mechanism of each prepared nano composites showed a good fit to PSORE model and Freundlich isotherm model. A reusability study was also conducted. The regeneration study demonstrated the reusability of the prepared beads after multiple treatments with 0.5 M HNO3 solution. These findings underscore the potential of alginate-based nanocomposites as sustainable, cost-effective solutions for heavy metal remediation, with promising implications for industrial water treatment applications.

احمد راشد صبري القصبي

اسلام سمير حسن

أحمد إبراهيم الحطاب

أحمد محمد عبد الله

أحمد على الشرقاوى

الهندسة

Rehabilitation of Flexible and Rigid Pavement: A Review

Road rehabilitation is crucial in maintaining and improving any country’s infrastructure. Roads are the helping hands that link communities, aid trade, and enable the movement of people and goods. This article mainly explores the rehabilitation of flexible and rigid pavement using asphalt and concrete overlays. It delves into existing knowledge gaps and suggests directions for future research. Synthesis literature was collected from the Scopus database based on specific keywords. It was then investigated using bibliometric analysis to discuss the trends and influential contributions in road rehabilitation. After eyeballing, refined literature was comprehensively addressed based on three key parameters: the total cost of road per year (EUAC), target service life, and international roughness index (IRI). The refined literature undergoes qualitative and statistical analyses, which are then summarized. The qualitative analysis revealed a tilted distribution of studies against rigid pavement overlays with the need to include additional parameters. Using Welch’s t-test, a statistical analysis was conducted to determine the development of the performance of the key parameters. The findings revealed no statistical significance of the key parameters for flexible and rigid pavement overlays. However, the EUAC parameter was statistically significance in favor of flexible pavement over rigid pavement with a p-value less than 0.01, whereas the remaining parameters showed no statistical significance between flexible and rigid pavement. The study aims to assess the collected studies and prioritize future work based on the limitations of these studies. Finally, this study is the first to collectively concentrate on these key parameters within the road rehabilitation domain.

مصعب جمال على سليمان لواش

أ.د. محمد محمد الغندور

أ.د. أشرف اسماعيل الصباغ 

أ.د. طارق عبد المنعم شرف                                

الهندسة

Predicting The Capacity of Cold-Formed Steel Hollow Sections under Elevated Temperatures Using Deep Learning

Because of the complex interaction between local and global instability modes at high temperatures, predicting the strength of thin-walled columns in fire situations is challenging. This study explores the use of machine learning to assess the fire resistance of steel hollow sections with rectangular and square shapes. Initially, a reliable finite element model is employed to assess column behavior, generating a comprehensive dataset that encompasses a variety of cross-sections, slenderness ratios, and temperatures. This dataset forms the backbone for training and evaluating machine learning models, specifically Deep Neural Network (DNN), Extreme Gradient Boosting XGBoost, and Support Vector Regression (SVR). Among these, the DNN model showcased pronounced accuracy, particularly when tested on previously unseen data; notably, its application for predicting the buckling capacity of these specific hollow sections under elevated temperatures represents a pioneering approach not previously investigated. This deep learning methodology offers a significant advantage by drastically reducing the computational effort required to determine buckling capacities compared to traditional, time-intensive finite element modelling. An in-depth error analysis further elucidated the origins of prediction discrepancies, underscoring the necessity of recognizing the model’s boundaries.

م. آلاء احمد محمد يعقوب

أ.د. السيد جلال الغندور
أ.د. عماد السعيد البلتاجي

د. حسام الدين علي وفقي

الهندسة

Integrating BIM And Blockchain for Automated Building Code Compliance Checking System

Integrating BIM And Blockchain for Automated Building Code Compliance Checking System The fourth industry revolution (Industry 4.0) inspired Architecture, Engineering, and Construction industry (AEC) research efforts for utilizing the innovated technologies. Integrating technologies with Building Information Modelling (BIM) furthers the ability to automate several construction data management processes. The integrated representation of the physical and functional characteristics of a building enabled by BIM provides a computational environment for several processes such as automated code compliance checking (ACC). Code compliance checking is essential during the design and construction phases. Performing manual code compliance checking is a human error-prone, labor-intensive, and time-consuming process. Thus, Developing ACC systems is becoming increasingly crucial with consideration that some provisions still require human intervention. Furthermore, the ACC process risks manipulation and human data tampering. This research aims to develop an integrated blockchain-based framework for ACC workflow to improve liability, traceability, and data transparency

احمد جمال عطية حنيفه

محمد الشخيبى

محمد الغندور

طارق شرف

أشرف الصباغ

الهندسه

Axial compressive behavior and design of confined concrete-filled stiffened steel tubular short columns

The axial compressive performance and design of square confined concrete-filled cold-formed stiffened steel tubular (CFSST) short columns, confined with Perfobond Leister (PBL) plates were investigated in this study. A numerical analysis using software ABAQUS is employed to analyze the ultimate strength of the generated composite columns, focusing on effect of PBL plates and internal angles on load-bearing capacity and concrete confinement. A parametric analysis was performed on 53 CFSST columns, including different thickness of steel, strength of infill concrete, and outer cross-sectional dimensions. Results show that the inclusion of PBL plates improves the axial compressive capacity by enhancing concrete confinement and reducing lateral expansion. The study also compares the ultimate axial resistances derived from numerical analyses with design provisions from EC4, BS5400, and DBJ 13-15 codes, demonstrating that the DBJ 13-15 code provides the closest predictions to numerical results. These findings suggest that PBL-confined cold-formed CFST columns offer a promising solution for structural applications requiring enhanced performance under axial compression.

عباد الرحمن محمد حسن

طارق شرف

محمد الغندور

أشرف الصباغ

الهندسة

Developing an Efficient Beam-Column Element for Advanced Analysis of Spatial Steel Frames

A novel Beam-Column element is present for advanced analysis of steel frames. The proposed model tries to overcome many issues of the analysis. A one element per member concept is used. Large deflection effects are considered by employing fifth order displacement function that takes member lateral loads into consideration. The interaction between the axial force effect and flexural actions is taken into consideration including the actions due to span loads. Plasticity spread along the member span is traced. A finite element program is prepared based on the proposed model and the accuracy has been assured.