Nour Eldeen M. Khalifa

Deep learning proved its efficiency in many fields of computer science such as computer vision, image classifications, object detection, image segmentation, and more. Deep learning models primarily depend on the availability of huge datasets. Without the existence of many images in datasets, different deep learning models will not be able to learn and produce accurate models. Unfortunately, several fields don't have access to large amounts of evidence, such as medical image processing. For example. The world is suffering from the lack of COVID-19 virus datasets, and there is no benchmark dataset from the beginning of 2020. This pandemic was the main motivation of this survey to deliver and discuss the current image data augmentation techniques which can be used to increase the number of images. In this paper, a survey of data augmentation for digital images in deep learning will be presented. The study begins and with the introduction section, which reflects the importance of data augmentation in general. The classical image data augmentation taxonomy and photometric transformation will be presented in the second section. The third section will illustrate the deep learning image data augmentation. Finally, the fourth section will survey the state of the art of using image data augmentation techniques in the different deep learning research and application.

Visual Question Answering is an extremely active research area in which the computer is given an image, a question in natural language, and it is required to give a correct answer to the question according to the semantics of the input image. The ability of VQA system to answer new questions about unseen images during training process is one main measure of effectiveness of the VQA model and this capability is called Zero-Shot VQA, but VQA datasets suffer from some problems that hinder good evaluation on models trained on these datasets. Firstly, Testing instances are not chosen perfectly to address how much the trained model accomplish the task of asking about new concepts that is not presented during training process. Secondly, most of visual question answering datasets suffer from problems in their contents such as small dataset size, leakiness of explicitly defined question types, and question types have abused evaluation scores that makes it difficult to evaluate algorithms on them. So models are not perfectly evaluated on such datasets. In order to avoid those evaluation obstacles, experiment is done on TDIUC dataset which has explicitly defined 12 question types, data are redistributed for zero shot task by re-splitting it to new training, val, and test instances such that test instances contains new concepts that is not presented in training data. Evaluation is done using methods that give a more representative measure of accuracy over all question types( Simple Accuracy, AMPT, HMPT) and one more evaluation schema(GMPT) is proposed for evaluating accuracy which is more expressive. Experiment shows that evaluation results on TDIUC dataset before redistributing train, val, and test sets for Zero Shot purpose gives inaccurate indicator of model performance (around 20% higher performance)

Coronavirus, also known as COVID-19, has spread to several countries around the world. It was announced as a pandemic disease by The World Health Organization (WHO) in 2020 for its devastating impact on humans. With the advancements in computer science algorithms, the detection of this type of virus in the early stages is urgently needed for the fast recovery of patients. In this paper, a study of neutrosophic set significance on deep transfer learning models will be presented. The study will be conducted over a limited COVID-19 x-ray. The study relies on neutrosophic set and theory to convert the medical images from the grayscale spatial domain to the neutrosophic domain. The neutrosophic domain consists of three types of images, and they are the True (T) images, the Indeterminacy (I) images, and the Falsity (F) images. The dataset used in this research has been collected from different sources. The dataset is classified into four classes {COVID-19, normal, pneumonia bacterial, and pneumonia virus}. This study aims to review the effect of neutrosophic sets on deep transfer learning models. The selected deep learning models in this study are Alexnet, Googlenet, and Restnet18. Those models are selected as they have a small number of layers on their architectures. To test the performance of the conversion to the neutrosophic domain, more than 36 trials have been conducted and recorded. A combination of training and testing strategies by splitting the dataset into (90-10%, 80-20%, 70-30) is included in the experiments. Four domains of images are tested, and they are, the original domain, the True (T) domain, the Indeterminacy (I) domain, and the Falsity (F) domain. The four domains with the different training and testing strategies were tested using the selected deep transfer models. According to the experimental results, the Indeterminacy (I) neutrosophic domain achieves the highest accuracy possible with 87.1% in the testing accuracy and performance metrics such as Precision, Recall, and F1 Score. The study concludes that using the neutrosophic set with deep learning models may be an encouraging transition to achieve better testing accuracy, especially with limited COVID-19 datasets.

The Coronavirus disease 2019 (COVID-19) is the fastest transmittable virus caused by severe acute respiratory syndrome Coronavirus 2 (SARS-CoV-2). The detection of COVID-19 using artificial intelligence techniques and especially deep learning will help to detect this virus in early stages which will reflect in increasing the opportunities of fast recovery of patients worldwide. This will lead to release the pressure off the healthcare system around the world. In this research, classical data augmentation techniques along with Conditional Generative Adversarial Nets (CGAN) based on a deep transfer learning model for COVID-19 detection in chest CT scan images will be presented. The limited benchmark datasets for COVID-19 especially in chest CT images are the main motivation of this research. The main idea is to collect all the possible images for COVID-19 that exists until the very writing of this research and use the classical data augmentations along with CGAN to generate more images to help in the detection of the COVID-19. In this study, five different deep convolutional neural network-based models (AlexNet, VGGNet16, VGGNet19, GoogleNet, and ResNet50) have been selected for the investigation to detect the Coronavirus-infected patient using chest CT radiographs digital images. The classical data augmentations along with CGAN improve the performance of classification in all selected deep transfer models. The outcomes show that ResNet50 is the most appropriate deep learning model to detect the COVID-19 from limited chest CT dataset using the classical data augmentation with testing accuracy of 82.91%, sensitivity 77.66%, and specificity of 87.62%.

Introduction: One attractive research area in the computer science field is soft biometrics. Aim: To
Identify a person’s gender from an iris image when such identification is related to security surveillance
systems and forensics applications. Methods: In this paper, a robust iris gender-identification method
based on a deep convolutional neural network is introduced. The proposed architecture segments
the iris from a background image using the graph-cut segmentation technique. The proposed model
contains 16 subsequent layers; three are convolutional layers for feature extraction with different
convolution window sizes, followed by three fully connected layers for classification. Results: The
original dataset consists of 3,000 images, 1,500 images for men and 1,500 images for women. The
augmentation techniques adopted in this research overcome the overfitting problem and make the
proposed architecture more robust and immune from simply memorizing the training data. In addition,
the augmentation process not only increased the number of dataset images to 9,000 images for the
training phase, 3,000 images for the testing phase and 3,000 images for the verification phase but
also led to a significant improvement in testing accuracy, where the proposed architecture achieved
98.88%. A comparison is presented in which the testing accuracy of the proposed approach was
compared with the testing accuracy of other related works using the same dataset. Conclusion: The
proposed architecture outperformed the other related works in terms of testing accuracy.

In this paper, a proposed system for automatic counting and visual multi-tracking for human sperm in microscopic video frames is presented. It can be easily turned into a commercial computer-assisted sperm analysis (CASA) system. CASA systems help in detecting infertility in human sperm according to clinical parameters. The proposed system consists of nine phases and it counts sperm in every single frame of video in real time and calculates the average sperm count through the whole video with accuracy 94.3% if it is compared to the manual counting. Also, it tracks all identified sperm in video frames in real time. It works with different frame rates above 15 frame/s to track visually the movements of the sperm. The dataset consists of three high-quality 1080p videos with different frame rates and durations. Finally, the open challenging research points are addressed.

In this paper, a system for aquarium family fish species identification is proposed. It identifies eight family fish species along with 191 sub-species. The proposed system is built using deep convolutional neural networks (CNN). It consists of four layers, two convolutional and two fully connected layers. A comparative result is presented against other CNN architectures such as AlexNet and VggNet according to four parameters (number of convolution and fully connected layers, the number of epochs in training phase to achieve 100% accuracy, validation accuracy, and testing accuracy). Through the paper, it is proven that the proposed system has competitive results against the other architectures. It achieved 85.59% testing accuracy while AlexNet achieves 85.41% over untrained benchmark dataset. Moreover, the proposed system has less trained images, less memory, less computational complexity in training, validation, and testing phases.

Abstract: A Blockchain is considered the main mechanism for Bitcoin currency. A Blockchain is known by a public ledger and public transactions stored in a chain. The properties of blockchain demonstrate in decentralization as distribution blocks, stability, anonymity, and auditing. Blockchain can enhance the results of network efficiency and improve the security of network. It also can be applied in several fields like financial and banking services, healthcare systems, and public services. However, the research is still opening at this point. It includes a big number of technical challenges which prevents the wide application of blockchain, for example, scalability problem, privacy leakage, etc. This paper shows a proposed comprehensive study of blockchain technology. It also examines the research efforts in blockchain. It presents a proposed blockchain lifecycle which refers to an evolution and a linked ring between business process management improvement and Internet-of-Things concepts. Then, this paper presents a practical proof of this relationship for smart city. It presents a new algorithm and a proposed blockchain framework for 38 blocks (which recognized as smart-houses). Finally, the future directions are well presented in blockchain field.

This paper is an extended version of "Deep Galaxy: Classification of Galaxies based on Deep Convolutional Neural Networks". In this paper, a robust deep convolutional neural network architecture for galaxy morphology classification is presented. A galaxy can be classified based on its features into one of three categories (Elliptical, Spiral, or Irregular) according to the Hubble galaxy morphology classification from 1926. The proposed convolutional neural network architecture consists of 8 layers, including one main convolutional layer for feature ex-traction with 96 filters and two principle fully connected layers for classification. The architecture is trained over 4238 images and achieved a 97.772% testing accuracy. In this version, "Deep Galaxy V2", an augmentation process is applied to the training data to overcome the overfitting problem and make the proposed architecture more robust and immune to memorizing the training data. A comparative result is present, and the testing accuracy was compared with those of other related works. The proposed architecture outperformed the other related works in terms of its testing accuracy.

Within the developments progress in Egypt today, the Egyptian student in the university should question himself about his role is in the advances in scientific and industrial revolutions. The university education in Egypt is still adopting traditional teaching methods which are incompatible with modern life, the student's and the lecturer mentality in the era of technology and development.

This paper presents a modern concept of nontraditional teaching methods and a realistic vision which is not only relying on e-learning in its current characteristics, but with the addition of some modern educational features such as self-learning. The most important features of self-learning are that it reduces the responsibilities of the lecturer and ensures that the students understand all the concepts and objectives of the courses. With the process of linking the courses in the Egyptian universities with the courses in the international universities (top
100 universities in the world), this will lead to a better education quality and better ranking of Egyptian universities. The pros and cons of the proposed idea will be discussed at the end of the research.

Abstract: Evidence-based medicine (EBM) is delivered through a didactic, blended learning, and mixed models. Students are supposed to construct an answerable question in PICO (patient, intervention, comparison, and outcome) framework, acquire evidence through search of literature, appraise evidence, apply it to the clinical case scenario, and assess the evidence in relation to clinical context.

Yet these teaching models have limitations especially those related to group work, for example, handling uncooperative students, students who fail to contribute, students who domineer, students who have personal conflict, their impact upon progress of their groups, and inconsistent individual acquisition of required skills.

At Pediatrics Department, Faculty of Medicine, Cairo University, we designed a novel undergraduate pediatric EBM assignment online system to overcome shortcomings of previous didactic method and aimed to assess its effectiveness by prospective follow-up during academic years 2012 to 2013 and 2013 to 2014.

The novel web-based online interactive system was tailored to provide sequential single and group assignments for each student. Single assignment addressed a specific case scenario question, while group assignment was teamwork that addressed different questions of same case scenario. Assignment comprised scholar content and skills.

We objectively analyzed students’ performance by criterion-based assessment and subjectively by anonymous student questionnaire.

A total of 2879 were enrolled in 5th year Pediatrics Course consecutively, of them 2779 (96.5%) logged in and 2554 (88.7%) submitted their work. They were randomly assigned to 292 groups. A total of 2277 (89.15%) achieved ≥80% of total mark (4/5), of them 717 (28.1%) achieved a full mark. A total of 2178 (85.27%) and 2359 (92.36%) made evidence-based conclusions and recommendations in single and group assignment, respectively (P < 0.001). A total of 1102 (43.1%) answered student questionnaire, of them 898 (81.48%) found e-educational experience satisfactory, 175 (15.88%) disagreed, and 29 (2.6%) could not decide. A total of 964 (87.47%) found single assignment educational, 913 (82.84%) found group assignment educational, and 794 (72.3%) enjoyed it.

Web-based online interactive undergraduate EBM assignment was found effective in teaching medical students and assured individual student acquisition of concepts and skills of pediatric EMB. It was effective in mass education, data collection, and storage essential for system and student assessment.

Wireless sensor networks (WSNs) have become an important component of our daily life. In the near future, they will dominate the technology industry around the world. WSNs have gained importance due to the variety of vital applications they can participate in, such as applications for the military, health care, agriculture, surveillance, and monitoring natural phenomena. Wireless multimedia sensor networks (WMSNs) are a special type of WSN.

Wireless Wireless Sensor Networks (WSNs) have become an important component in our daily lives. In near future, it will dominate the technology industry around the world. WSNs gain its importance due to the variety of vital applications it can participate in such as military, health care, agriculture, surveillance and monitoring natural phenomena applications.WSNs consist of small devices with limited energy and storage capabilities, called sensor nodes. The sensor nodes collect data from physical or environmental phenomena. They cooperatively pass the sensed data through the network to a certain location or sink node where the data can be collected and analyzed.Due to the unprotected nature of wireless communication channels and untrusted transmission medium of WSNs, it becomes vulnerable to many types of security attacks. The attackers ultimately seek to eavesdrop, steal confidential data, injecting false data or even jamming the whole network, so securing these networks becomes a must.In this paper, a proposed security schema for WSNs will be introduced. The proposed security schema will be appropriate for real time multimedia streaming. It will construct its security features within the application and transport layer as the information that the attackers seek ultimately exist within these layers.The proposed security schema consists of two security levels; the first level is encrypting the packet data using Advanced Encryption Standard (AES) while the second level is generating Message Authentication Code (MAC) using Cipher-based Message Authentication Code (CMAC). Both levels achieved the principles of WSNs security and they are (authentication, confidentiality, data integrity and availability).Performance comparisons between the proposed security schema and other security frameworks are presented. Finally, all the presented work in this research was developed and implemented using Network Simulator-2 (NS-2). According to our literature reviews, this research is one of the first researches that use NS-2 as a security simulator. As NS-2 does not support any security features before

Wireless sensor networks (WSNs) are usually deployed for gathering data from unattended or hostile environments. Therefore, securing data transmission across these environments is a must. Due to the fact that the sensors have a limited power, any security mechanism for sensor network must be energy efficient. In this paper, a secure energy efficient mechanism is introduced with a proposed scenario which leads to a significant improvement in network energy consumption. The mechanism constructs its security features in the application and transport layer as the information that the attackers seek ultimately resides within these layers. We modified the packet format for WSN. Data payload was encrypted by Advanced Encryption Standard (AES) and Message Authentication Code (MAC) was generated to assure data confidentiality and integrity. The energy consumption metric has been taken into considerations while designing and testing the mechanism to make it energy efficient as much as possible. The energy efficiency was achieved by giving a higher priority to the secured packet over the normal packet in the Interface Queue (IFQ). Through this paper, a detailed structure of the proposed mechanism is introduced and implemented using Network Simulator-2 (NS-2). This is the first research that implements security algorithms within NS-2. Since NS-2 does not support any security features before, this research will be a good start to begin using NS-2 as a security simulator.

The nature of playing video streams over a network in a real time requires that the transmitted frames are sent in a bounded delay. Also, video frames need to be displayed at a certain rate; therefore, sending and receiving encrypted packets must be sent in a certain amount of time in order to utilize the admissible delay. There are two main components for securing multimedia transmissions. Key management protocols such as transport layer security (TLS) protocol is used as cryptographic protocol that provides security and data integrity for communications over IP networks. Advanced encryption standard (AES) is used as a symmetric block cipher, that can encrypt (encipher) and decrypt (decipher) video frames. The real-time transport protocol (RTP) protocol is used as a standardized packet format for delivering video over the IP Networks. This makes secure video encryption feasible for real-time applications without any extra dedicated hard-ware. In this paper; we studied the impact of frame rate on a secured real time transmission of video over IP networks. Through this paper, we will try to find the most suitable frame rate in order to achieve better data rate and fewer frame and packet loss. This is done via java as a programming language for implementation.

The world has become a small village. The internet has played an important role in the communications revolution. It facilitates the management of institutions and organizations in many countries. Due to advances in using multimedia streaming (especially video), it becomes important to secure real time video transmission.
Securing video depends on many factors. The key two factors are: the complexity of encryption algorithm and time to break the encryption by hackers or competitors.
There are two main components for securing multimedia transmissions: security protocols and text encryption algorithms. Security protocols have been used to provide security and data integrity for communications over IP networks. Text encryption algorithms have been used for encrypting and decrypting video frames.
This thesis proposed a security scheme for streamed video frames over IP network. We used Datagram Transport Layer Security (DTLS) protocol. We applied the most four famous text encryption algorithms on video frames. These algorithms are Advanced Encryption Standard (AES), Data Encryption Standard (DES), Triple Data Encryption Standard (3DES) and Blowfish. They were used to determine the most suitable one to be selected to secure video frames. We compared our results with previous results on video frames encryption. We also modified the Real-time Transport Protocol (RTP) to adapt the proposed security scheme. RTP is used as a standardized packet format for delivering video over the IP networks. Our proposed scheme would be feasible for real time applications without any extra dedicated hardware.
After determining the most suitable encryption algorithm, we compared the proposed security scheme with normal video transmission scheme according to the following parameters: Frame Loss, Packet Loss and Data Rate. We evaluated the overheads of the proposed scheme. We also studied the impact of frame rate on securing real time video transmission over IP networks.
We used JAVA programming language to develop the proposed scheme. We found that the secured connection setup needed more time than normal one. The secured setup time is an offline process, so it does not affect the transmission of video frames in the real time. Using random generated keys for encryption process for every connection makes attacks harder to happen. The AES algorithm is considered the most suitable text encryption algorithm to be used in video, as it achieves less encryption time per frame. We found that the size of the overhead was very small compared to video original size.
Finally, we studied the impact of frame rate on securing real time video streaming. We determined the best frame rate needed to achieve fewer frame, packet loss and best quality on video.