Future University In Egypt (FUE)
Future University is one of most promising private universities in Egypt. Through excellence in teaching, research and service, Future University strives to provide a comprehensive, high-quality education that prepares our graduates to be future leaders.
mainLogo
90th Street
New Cairo
Egypt

Mohamed Ismail Roushdy

Basic information

Name : Mohamed Ismail Roushdy
Title: Prof. Dr.
Google Schoolar Link
Personal Info: Professor & Dean of faculty of Computers & Information Technology, Future University in Egypt

Education

Certificate Major University Year
PhD . Ain Shams University - Faculty of Science 1993
Masters . . 1984
Bachelor . . 1979

Researches /Publications

Using K-Nearest Neighbors and Support Vector Machine Classifiers in Personal Identification based on EEG Signals - 01/0

Mohamed Ismail Roushdy

Shaymaa adnan Abdulrahman ,Abdel-Badeeh M. Salem

01/05/2020

In the present paper, electroencephalogram (EEG) data have been used to human identification by computing sample entropy and graph entropy as feature extractions. Used two classifier types , which are K-Nearest Neighbors (K-NN) and Support Vector Machine (SVM). Python and Matlab software were used in this study and EEG data was collected by UCI repository . Matlab used when Thirteen channels was applied as feature extraction . The experimental results show that, Python software classifies the EEG-UCI data better than MATLAB environment software where the accuracy of KNN and SVM were 85.2% and 91.5% respectively.

Download PDF
Overview of Acquisition Techniques Brain Signals in Human Identification and Disease Diagnosis: Applications and Challenges - 01/0

Mohamed Ismail Roushdy

Shaymaa Adnan Abdulrahman,Abdel-Badeeh M. Salem

01/05/2020

Electroencephalogram (EEG) signals refer to distinctive neurons’ electrical activity, depiction that upkeep biometric recognition. Usually in biometrics, the acquisition protocol has been important for EEG-based biometric system performance. Various acquisition protocols brain signals like evoked potentials besides relaxation, motor and non-motor imaginary have shown and discussed. This study discusses the potentials for identifying an individual based on EEG signals and highpoint the challenges of employing brain signals as a biometric modality in disease identification. Also, to discuss diverse solutions for limiting and decreasing their effects . Finally, an overview of EEG biometrics investigations has presented with findings and conclusions. Through this study it found that sensor like electrodes is help to disease diagnosis to detection disorder of the brain or increased power of the lower frequency bands and a decrease of high frequencies of patient's brain

Download PDF
Comparative study for 8 computational intelligence algorithms for human identification - 01/0

Mohamed Ismail Roushdy

Shaymaa Adnan Abdulrahman , Wael Khalifa, Abdel-Badeeh ,M. Salem

01/05/2020

The biometric system includes the algorithms, procedures, and devices which are utilized for the purpose of recognizing individuals according to their behavioral and physiological features. The approaches of Computational Intelligence (CI) are utilized extensively to establish biometric-based identities as well as overcoming non-idealities usually exist in samples. The objective of this paper is to analyze and evaluate the various computational intelligence (CI) approaches for the human identification based on biometrics . The study includes 8 top CI algorithms, namely; k-Nearest Neighbor(K-NN), Artificial Neural Networks (ANNs), Support vector machines (SVMs), Fuzzy Discernibility Matrix (FDM), Naïve Bayes (NB), k-means, Decision Trees (DTs), and Genetic algorithms (GAs). Also the study provides the technical characteristics and features of these algorithms as well as finds advantages and disadvantages of these methods . The analyzed algorithms can be selected according to quantity and quality of data presented at work.

Download PDF
Reversible Watermarking for Protecting Patient’s Data Privacy Using an EPR-Generated QR Code - 01/0

Mohamed Ismail Roushdy

Alaa H. ElSaadawy, Ahmed S. ELSayed, M. N. Al-Berry

01/03/2020

As a result of the modern communication technology and the transmission of medical images from one place to another, watermarking has become urgent to protect copyright, authentication and integrity of medical images and patient’s information. This paper proposes a solution to the authentication and integrity problems by proposing a fragile reversible watermarking scheme. The proposed scheme generates a Quick Response (QR) code from the patient’s data and uses it as a watermark hidden in the medical image. The proposed scheme resulted in an average Peak Signal to Noise Ratio (PSNR) of 51.25 dB for a payload of one bit per pixel (1 bpp), Structural Similarity Index Measure (SSIM) of 0.99 and Mean Squared Error (MSE) of 0.488 indicating that the watermarked images obtained are of high visual quality.

Download PDF
Metadata Extraction for Low-Quality Semi-structured Spreadsheets - 01/0

Mohamed Ismail Roushdy

Arwa Awad,Rania Elgohary,Ibrahim Moawad

01/03/2020

The approach in which data is stored in Excel spreadsheets is not the most ideal approach to sort out and access it. Despite using a spreadsheet is in a growing exponentially in the last years, the increases in volume and complexity of these data have led to increased requirements to preserve these data and reuse it. Most of the applications focused on extracting structured data from semi-structured data on the web by HTML and XML formats. Nowadays, there is an explosion of semi-structured documents that generate outside the web. Metadata is important to extract the spreadsheet data. This paper produced an automated relation extractor tool that lets ordinary users extract structured relational tables from spreadsheets without previous experience. To define a structure for the spreadsheet table, this paper introduces a framework that automatically extracts relational data (tables) from spreadsheets and converts from Low-Quality data to High-Quality data. The paper considers some perspective approaches for the clustering-based table detection, heuristic-based table metadata extraction and rule-based table analysis for attribute names to automate insertion into a structured database to facilitate, integrate and reuse the data stored in spreadsheets.

Download PDF
SUPPORT VECTOR MACHINE APPROACH FOR HUMAN IDENTIFICATION BASED ON EEG SIGNALS - 01/0

Mohamed Ismail Roushdy

Shaymaa adnan Abdulrahman, Abdel-Badeeh M. Salem

01/02/2020

The signals of the electroencephalogram (EEG) have been applied for detecting as well as registering the electrical efficiency in the human brain. In this paper, EEG signals have been utilized for human identification. The reliability regarding a lot of biometric systems aren’t adequate due to the possibility of being copied or faked. Thus the brain signatures have been applied as potential biometric identifiers. The aim of this paper is to apply sample entropy and graph entropy as feature extraction. While in classification Support vector machine (SVM) and KNearest Neighbor (KNN) have achieved. Machine Learning Repository (UCI) used as dataset. Experimental consequences on this dataset demonstrate substantial enhancement in the classification accuracy as compared with other testified results in the literature. Results showed that the classification accuracy with SVM for biometric identification is 90.8% while with K-NN is 83.7% .Our study using13channels to feature extraction.

Download PDF
Impact of segmentation on iris liveness detection - 01/1

Mohamed Ismail Roushdy

Manar Ramzy Dronky ,Wael Khalifa

01/12/2019

Applying iris recognition systems in many sensitive security areas highlights the importance of developing liveness detection methods. These methods read the users physiological signs of life to verify if the iris pattern acquired for identification is fake or real. This paper explores the results of BSIF for solving the problem of iris liveness detection to combat presentation attacks. Four public datasets representing printed, plastic, synthetic and contact lens attacks were used for method evaluation in both scenarios segmented and unsegmented eye images. The results have showed that BSIF can efficiently detect plastic and synthetic attacks without segmentation with correct classification rate of 100%. In addition, unsegmented eye images achieved better results in detecting print attack on the tested datasets. While, segmentation is still required in the most challenging attack which is by contact lens.

Download PDF
Lung Nodule Detection and Classification using Random Forest: A Review - 01/1

Mohamed Ismail Roushdy

Nada S. El-Askary,Mohammed A.-M. Salem

01/12/2019

Lung nodule is an abnormal growth of tissues in the lung that can be an onset for lung cancer. Fast detection for those nodules and classifying them will ensure better chances for treatments. Random Forest (RF) is a powerful machine learning algorithm and a state-of-the-art technology that proved to give rewarding results in helping radiologies diagnosing lung pathologies. The paper presents a survey on recent researches made for lung nodule detection and classification using RF. Wide range of datasets can be used for lung nodule detection are listed. Different models with the used features and their results are discussed in this review.

Download PDF
Hybrid Method for Modeling User Interests based on Social Network - 01/1

Mohamed Ismail Roushdy

Marina Shafik,Rania Elgohary,Ibrahim Moawad

01/12/2019

The growing popularity of social networks and microblogging services has gradually increased the demand for personalized applications. The microblogging services such as Twitter has a powerful forum for users to share their personal interest and opinions. Mining and analyzing user's interests is a crucial factor in buying decisions and tracking the emotions of the public about their items, business, etc. Although, Twitter has a broad range of topics in real-time, it poses significant challenges because of the unstructured short text. In this paper, the best model for finding the user's topics of interest is being investigated by building the profile of individual users based on their tweets. A hybrid Topic-based model is proposed that combines both two unsupervised learning algorithms with sentiment consideration and user features. Thus, we show that the proposed hybrid model has a higher performance in the topic extraction of user's interests on Social Networks.

Download PDF
A Review on Iris Liveness Detection Techniques - 01/1

Mohamed Ismail Roushdy

Manar Ramzy Dronky,Wael Khalifa

01/12/2019

Iris recognition systems have been widely deployed for authentication in many sensitive security areas for their accuracy and consistency. However, as the iris technology evolves, ways to attack it evolve too. Fake iris samples could be used to spoof the iris recognition system. As a result, Iris liveness detection methods have been developed. These methods read the users physiological signs of life to verify if the iris pattern acquired for identification is fake or real. In this paper, a review for the previous work done in iris liveness detection is presented.

Download PDF
Multi-view Convolutional Neural Network for lung nodule false positive reduction - 01/1

Mohamed Ismail Roushdy

Salsabil Amin El-Regaily,Mohammed Abdel Megeed Salem,Mohamed Hassan Abdel

01/10/2019

Background and objective: Computer-Aided Detection (CAD) systems save radiologists time and provide a second opinion in detecting lung cancer by performing automated analysis of the scans. False positive reduction is one of the most crucial components of these systems that play a great role in the early diagnosis and treatment process. The objective of this paper is to efficiently handle this problem by detecting nodules and separating them from a large number of false positive candidates. Methods: The proposed algorithm segments lungs and nodules through a combination of 2D and 3D region growing, thresholding and morphological operations. Vessels and most of the internal lung structure have a tabular shape that differs from the compact rounded shape of nodules, therefore they are eliminated by building and thresholding a3D depth map, to produce the initial candidates. To reduce the number of false positives, a rule-based classifier is used to eliminate the obvious non-nodules, followed by a multi-view Convolutional Neural Network. The convolutional network is built specifically to handle the provided inputs and is customized to provide the best possible outputs without the extra computational complexity that is required when compared to a 3D network. 650 cases from the LIDC dataset are used to train and test the network. For each candidate, the axial, coronal and sagittal views are extracted and fed to the three network streams. Results: The proposed algorithm achieved a high detection sensitivity of 85.256%, a specificity of 90.658% and an accuracy of 89.895%. Experimental results indicate that the proposed algorithm outperforms most of the other algorithms in terms of accuracy and sensitivity. The proposed solution achieves a good tradeoff between efficiency and effectivity and saves much computation time. Conclusion: The work shows that the proposed multi-view 2D network is a simple, yet effective algorithm for the false positive reduction problem. It can detect nodules that are isolated, linked to a vessel or attached to the lung wall. The network can be improved to detect ground glass nodules in the future.

Download PDF

Awards

Award Donor Date
University Appreciation Award in Technological Sciences Ain Shams University 2018

Follow us on

Visit the Faculty

ADS