Mohammed Kayed

Followers

Following

Co-author

Public Views

Interests

Uploads

Papers by Mohammed Kayed

A novel and accurate deep learning-based Covid-19 diagnostic model for heart patients

Signal, Image and Video Processing, May 19, 2023

Using radiographic changes of COVID-19 in the medical images, artificial intelligence techniques ... more Using radiographic changes of COVID-19 in the medical images, artificial intelligence techniques such as deep learning are used to extract some graphical features of COVID-19 and present a Covid-19 diagnostic tool. Differently from previous works that focus on using deep learning to analyze CT scans or X-ray images, this paper uses deep learning to scan electro diagram (ECG) images to diagnose Covid-19. Covid-19 patients with heart disease are the most people exposed to violent symptoms of Covid-19 and death. This shows that there is a special, unclear relation (until now) and parameters between covid-19 and heart disease. So, as previous works, using a general diagnostic model to detect covid-19 from all patients, based on the same rules, is not accurate as we prove later in the practical section of our paper because the model faces dispersion in the data during the training process. So, this paper aims to propose a novel model that focuses on diagnosing accurately Covid-19 for heart patients only to increase the accuracy and to reduce the waiting time of a heart patient to perform a covid-19 diagnosis. Also, we handle the only one existed dataset that contains ECGs of Covid-19 patients and produce a new version, with the help of a heart diseases expert, which consists of two classes: ECGs of heart patients with positive Covid-19 and ECGs of heart patients with negative Covid-19 cases. This dataset will help medical experts and data scientists to study the relation between Covid-19 and heart patients. We achieve overall accuracy, sensitivity and specificity 99.1%, 99% and 100%, respectively. Keywords Coronavirus • Electro diagrams • Deep learning • Heart patients 1 Introduction Covid-19 is a health disaster wherein there are 43 million Covid-19 positive cases and 1.2 million people have died as a result. It is necessary to develop an automatic, early and accurate COVID-19 diagnostic mechanism. The disease is typically detected using reverse-transcription polymerase chain reaction (RT-PCR) testing. In spite of this, RT-PCR has been found that the sensitivity of it is not high enough for early detection of COVID-19. Also, the supply of kits of RT-PCR is different from a country to another and many

Download

Ensemble Federated Learning: An approach for collaborative pneumonia diagnosis

Applied Soft Computing

Pneumonia Detection on Chest X-ray Images Using Ensemble of Deep Convolutional Neural Networks

Applied Sciences

Pneumonia is a life-threatening lung infection resulting from several different viral infections.... more Pneumonia is a life-threatening lung infection resulting from several different viral infections. Identifying and treating pneumonia on chest X-ray images can be difficult due to its similarity to other pulmonary diseases. Thus, the existing methods for predicting pneumonia cannot attain substantial levels of accuracy. This paper presents a computer-aided classification of pneumonia, coined Ensemble Learning (EL), to simplify the diagnosis process on chest X-ray images. Our proposal is based on Convolutional Neural Network (CNN) models, which are pretrained CNN models that have been recently employed to enhance the performance of many medical tasks instead of training CNN models from scratch. We propose to use three well-known CNNs (DenseNet169, MobileNetV2, and Vision Transformer) pretrained using the ImageNet database. These models are trained on the chest X-ray data set using fine-tuning. Finally, the results are obtained by combining the extracted features from these three model...

Download

A novel and accurate deep learning based Covid-19 diagnostic model for heart patients only

Download

Dental Age Estimation Using Deep Learning: A Comparative Survey

Computation

The significance of age estimation arises from its applications in various fields, such as forens... more The significance of age estimation arises from its applications in various fields, such as forensics, criminal investigation, and illegal immigration. Due to the increased importance of age estimation, this area of study requires more investigation and development. Several methods for age estimation using biometrics traits, such as the face, teeth, bones, and voice. Among then, teeth are quite convenient since they are resistant and durable and are subject to several changes from childhood to birth that can be used to derive age. In this paper, we summarize the common biometrics traits for age estimation and how this information has been used in previous research studies for age estimation. We have paid special attention to traditional machine learning methods and deep learning approaches used for dental age estimation. Thus, we summarized the advances in convolutional neural network (CNN) models to estimate dental age from radiological images, such as 3D cone-beam computed tomograp...

Download

Hierarchical cloud architecture for identifying the bite of “Egyptian cobra” based on deep learning and quantum particle swarm optimization

Scientific Reports, Mar 31, 2023

One of the most dangerous snake species is the "Egyptian cobra" which can kill a man in only 15 m... more One of the most dangerous snake species is the "Egyptian cobra" which can kill a man in only 15 min. This paper uses deep learning techniques to identify the Egyptian cobra bite in an accurate manner based on an image of the marks of the bites. We build a dataset consisting of 500 images of cobra bites marks and 600 images of marks of other species of snakes that exist in Egypt. We utilize techniques such as multi-task learning, transfer learning and data augmentation to boost the generalization and accuracy of our model. We have achieved 90.9% of accuracy. We must keep the availability and accuracy of our model as much as possible. So, we utilize cloud and edge computing techniques to enhance the availability of our model. We have achieved 90.9% of accuracy, which is considered as an efficient result, not 100%, so it is normal for the system to perform sometimes wrong classifications. So, we suggest to retrain our model with the wrong predictions, whereas the edge computing units, where the classifier task is positioned, resend the wrong predictions to the cloud model, where the training process occurs, to retrain the model. This enhances the accuracy to the best level after a small period and increases the dataset size. We use the quantum particle swarm optimization technique to determine the optimal required number of edge nodes. According to the world health organization (WHO), there are 5 million persons are exposed to snake bites annually. This leads to occurring 2.5 million cases of poisoning, 137.880 deaths and also 300.000 cases of amputation and paralysis per year. So, snakes are among the most three dangerous animals to humans. Out of 3848 identified species of snakes, around 800 species are venomous, among which only about 50 species are fatal to humans. The "Egyptian cobra" is of the most poisonous snakes in my country "Egypt" and its venom is able to kill a man in only 15 min, and an elephant in 3 h 1,2. Egyptian cobras prefer to stay near the human houses to hunt domestic fowl like chickens. This increases its danger to the Egyptian people. Generally, one of the quickest required measures after a snake bite is to identify the species of the attacking snake to take the right antitoxin. In real situations, doctors mostly give polyvalent anti-venom to the snake bite victim without considering which snake species has bitten the victim. Even if the patient observed some visual features of the snake, the medical person is not expert in the snakes species to determine the correct species of the attacking snake from the description taken from the victim. Also, the victim in this situation is not aware enough to focus on snake features. In the case of the polyvalent anti-venom is given to the victim, polyvalent anti-venom includes antibodies against more than one snake species. So, polyvalent anti-venom may neutralize the venom of a single snake bite and the remaining non-neutralized part of the polyvalent anti-venom causes further risk to the human health 3. So proper identification of the snake is very important for the proper medical treatment to save the life of the snake bite victims 4. Computer vision technique has developed quickly in the area of automated object classification and recognition 5-8. Computer scientists use different machine learning techniques for image classification. Image classification, based on machine learning, contains two phases: feature extraction and classification. During training, the pre-defined classes (labels) can be conceived of the available dataset that extracts the characteristic features from each image class and forms a special description for each specific class. Using machine learning for

Download

A model for generating a user dynamic profile on social media

Journal of King Saud University - Computer and Information Sciences

Medical Image Classification Using Transfer Learning and Chaos Game Optimization on the Internet of Medical Things

Computational Intelligence and Neuroscience

The Internet of Medical Things (IoMT) has dramatically benefited medical professionals that patie... more The Internet of Medical Things (IoMT) has dramatically benefited medical professionals that patients and physicians can access from all regions. Although the automatic detection and prediction of diseases such as melanoma and leukemia is still being investigated and studied in IoMT, existing approaches are not able to achieve a high degree of efficiency. Thus, with a new approach that provides better results, patients would access the adequate treatments earlier and the death rate would be reduced. Therefore, this paper introduces an IoMT proposal for medical images’ classification that may be used anywhere, i.e., it is an ubiquitous approach. It was designed in two stages: first, we employ a transfer learning (TL)-based method for feature extraction, which is carried out using MobileNetV3; second, we use the chaos game optimization (CGO) for feature selection, with the aim of excluding unnecessary features and improving the performance, which is key in IoMT. Our methodology was eva...

Download

Sign language identification and recognition: A comparative study

Open Computer Science

Sign Language (SL) is the main language for handicapped and disabled people. Each country has its... more Sign Language (SL) is the main language for handicapped and disabled people. Each country has its own SL that is different from other countries. Each sign in a language is represented with variant hand gestures, body movements, and facial expressions. Researchers in this field aim to remove any obstacles that prevent the communication with deaf people by replacing all device-based techniques with vision-based techniques using Artificial Intelligence (AI) and Deep Learning. This article highlights two main SL processing tasks: Sign Language Recognition (SLR) and Sign Language Identification (SLID). The latter task is targeted to identify the signer language, while the former is aimed to translate the signer conversation into tokens (signs). The article addresses the most common datasets used in the literature for the two tasks (static and dynamic datasets that are collected from different corpora) with different contents including numerical, alphabets, words, and sentences from diffe...

Download

Review paper on “ Optimized approaches for web data harvesting

Companies spend millions of dollars to build data warehouses to hold their data and data mining a... more Companies spend millions of dollars to build data warehouses to hold their data and data mining applications must take advantage of this. Besides saving significant manual effort and storage space, relational integration allows data mining applications to access the most up-to-date information available. To collect large data is very tedious and frustrating task. There are various kinds of valuable semantic information about realworld entities embedded in web pages and databases. Extracting and integrating these entity information from the Web is of great significance.

Download

Extracting Geographic Addresses from Social Media using Deep Recurrent Neural Networks

2021 9th International Japan-Africa Conference on Electronics, Communications, and Computations (JAC-ECC), 2021

The importance of geographical, addresses in people's daily lives cannot be underestimated. P... more The importance of geographical, addresses in people's daily lives cannot be underestimated. People usually use the Internet to search for unfamiliar areas and then use map services to mark locations. Using social media to extract information, particularly geographical addresses, is rapidly increasing worldwide. Social media represents the right choice as a source in identifying the location that people need to find. In this paper, a deep neural network using a Bidirectional Long Short-Term Memory with CRF (BI-LSTM-CRF) model is applied for address extraction. In addition, a Bidirectional Encoder Representations from Transformers (BERT) model is implemented to extract the geographical addresses from Facebook posts. Further, we reveal how to use the BIEO tagging method to apply the sequence labeling technique to Arabic postal address extraction. An Arabic corpus from social media is annotated to evaluate our proposed model. The results show that Arabic postal addresses can be extracted through BI-LSTM-CRF and BERT models with a high F-measure.

Criteria for Evaluating Information Extraction Systems

The Internet presents a huge amount of useful information which is usually formatted for its user... more The Internet presents a huge amount of useful information which is usually formatted for its users, which makes it difficult to extract relevant data from various sources. Therefore, the availability of robust, flexible Information Extraction (IE) systems that transform the Web pages into program-friendly structures will become a great necessity. Although many approaches for data extraction from Web pages have been developed, there has been limited effort to compare such tools. In addition to briefly surveying the major data extraction approaches described in the literature, the paper also mainly presenting three classes of criteria for qualitatively analyzing these approaches. The criteria of the first class are concerned with the difficulties of an IE task, so these criteria are capable of determining why an IE system fails to handle some Web sites of particular structures. The criteria of the second class are concerned with the effort made by the user in the training process, so ...

A Comparative study for Feature Extraction and Classification of Images

2019 6th International Conference on Advanced Control Circuits and Systems (ACCS) & 2019 5th International Conference on New Paradigms in Electronics & information Technology (PEIT), 2019

Digital images increased rapidly due to the usage of users to the Internet, digital camera and th... more Digital images increased rapidly due to the usage of users to the Internet, digital camera and their cell phones. The huge amounts of images are used for different purposes: business, education, recording/sharing events, remembering, etc. Most commercial and online shopping companies rely on images to attract consumers to visit and buy their products. Image classification is a core task for most of such applications. The performance of a classifier depends on two main factors: the features extracted from the images and the classification techniques. Many feature extraction algorithms are proposed by researchers such as Local Binary Pattern, Maximally Stable Extermal Regions, Speed Up Robust Features and Deep Convolutional Neural Networks. This paper tries to experimentally compare among these different feature extraction algorithms by applying different classifiers with each one. These classifiers are Support Vector Machine (SVM), Decision Tree, k-Nearest Neighbors and Naive Bayes. The classifiers were learned from a dataset of eight categories: animals, food, nature, building, tourism, people, sport and apparatus. Our experiments show that SVM achieves 93.89% accuracy when Residual Neural Network is used as a feature detector and SVM as a classifier.

Classification of Garments from Fashion MNIST Dataset Using CNN LeNet-5 Architecture

2020 International Conference on Innovative Trends in Communication and Computer Engineering (ITCE), 2020

Recently, deep learning has been used extensively in a wide range of domains. A class of deep neu... more Recently, deep learning has been used extensively in a wide range of domains. A class of deep neural networks that give the most rigorous effects in solving real-world problems is a Convolutional Neural Network (CNN). Fashion businesses have used CNN on their e-commerce to solve many problems such as clothes recognition, clothes search and recommendation. A core step for all of these implementations is image classification. However, clothes classification is a challenge task as clothes have many properties, and the depth of clothes categorization is highly complicated. This complicated depth makes different classes to have very similar features, and so the classification problem becomes very hard. In this paper, CNN based LeNet-5 architecture is proposed to train parameters of the CNN on Fashion MNIST dataset. Experimental results show that LeNet-5 model achieved accuracy over 98%. Therefore, it outperforms both the classical CNN model and the other existing state-of-the-art models in literatures.

GenDE: A CRF-Based Data Extractor

Journal of Web Engineering, 2020

Web site schema detection and data extraction from the Deep Web have been studied a lot. Although... more Web site schema detection and data extraction from the Deep Web have been studied a lot. Although, few researches have focused on the more challenging jobs: wrapper verification or extractor generation. A wrapper verifier would check whether a new page from a site complies with the detected schema, and so the extractor will use the wrapper to get instances of the schema types. If the wrapper failed to work with the new page, a new wrapper/schema would be re-generated by calling an unsupervised wrapper induction system. In this paper, a new data extractor called GenDE is proposed. It verifies the site schema and extracts data from the Web pages using Conditional Random Fields (CRFs). The problem is solved by breaking down an observation sequence (a Web page) into simpler subsequences that will be labeled using CRF. Moreover, the system solves the problem of automatic data extraction from modern JavaScript sites in which data/schema are attached (on the client side) in a JSON format. ...

Download

FiVaTech2: A Supervised Approach to Role Di erentiation for Web Data Extraction From Template Pages (人工知能学会全国大会 (第 26 回) 文化, 科学技術と未来)--(International Organized Session 「Special Session on Web Intelligence & Data Mining」)

A Survey of Web Information Extraction Systems

IEEE Transactions on Knowledge and Data Engineering, 2006

The Internet presents a huge amount of useful information which is usually formatted for its user... more The Internet presents a huge amount of useful information which is usually formatted for its users, which makes it difficult to extract relevant data from various sources. Therefore, the availability of robust, flexible Information Extraction (IE) systems that transform the Web pages into program-friendly structures such as a relational database will become a great necessity. Although many approaches for data extraction from Web pages have been developed, there has been limited effort to compare such tools. Unfortunately, in only a few cases can the results generated by distinct tools be directly compared since the addressed extraction tasks are different. This paper surveys the major Web data extraction approaches and compares them in three dimensions: the task domain, the automation degree, and the techniques used. The criteria of the first dimension explain why an IE system fails to handle some Web sites of particular structures. The criteria of the second dimension classify IE systems based on the techniques used. The criteria of the third dimension measure the degree of automation for IE systems. We believe these criteria provide qualitatively measures to evaluate various IE approaches.

Download

Extending the lifetime of wireless sensor networks: A hybrid routing algorithm

Computer Communications, 2012

Wireless Sensor Network (WSN) is one of the underlying mechanics of IoT (Internet-of-Things) whic... more Wireless Sensor Network (WSN) is one of the underlying mechanics of IoT (Internet-of-Things) which has recently seen a worldwide interest by its use in several domains as military, automation, agriculture, environment, underwater and etc. Energy efficiency and reliability of transmitted messages are two of the major requirements for the surveillance and wireless detection applications in WSN. For this reason, a growing number of research studies are carried out. Several routing protocols have been developed to furnish better performance for optimizing the network's energy consumption in WSN; most of them are based in clustering and hierarchical topology. However, most of these routing protocols cannot consider all required and important metrics to increase the lifespan of the network. This work introduces an enhanced algorithm of the Low Energy Adaptive Clustering Hierarchy (LEACH) protocol called Balanced Current Energy-LEACH (BCE-LEACH) which aims to equity the power consumption of sensor nodes of the network due to extend the lifetime of the network. The proposed protocol is based on the residual energy of sensor nodes to pick out cluster heads (CHs) that means only nodes with sufficient current energy can participate in CH selection, and then it is focused on both current energy and the distance toward the base station (BS) to select a parent CH which has current energy greater than the mean energy of CHs and the distance to the BS is fewer than the average. This parent CH groups all data from other CHs. The root CH after doing data aggregation sends compressed information to the BS using the multi-hop process between CHs. The improved algorithm is simulated in MATLAB R2016b simulation tool. Simulation results in this paper indicate that the introduced protocol works better than LEACH through extending the lifetime of the WSN.

Download

Improving Crisis Events Detection Using DistilBERT with Hunger Games Search Algorithm

Mathematics

This paper presents an alternative event detection model based on the integration between the Dis... more This paper presents an alternative event detection model based on the integration between the DistilBERT and a new meta-heuristic technique named the Hunger Games Search (HGS). The DistilBERT aims to extract features from the text dataset, while a binary version of HGS is developed as a feature selection (FS) approach, which aims to remove the irrelevant features from those extracted. To assess the developed model, a set of experiments are conducted using a set of real-world datasets. In addition, we compared the binary HGS with a set of well-known FS algorithms, as well as the state-of-the-art event detection models. The comparison results show that the proposed model is superior to other methods in terms of performance measures.

Download

Bitcoin and Blockchain

Mohammed Kayed

Uploads

Papers by Mohammed Kayed

Log In