Accepted Papers



Biomedical Study of Demographics and Clinical Features of Lichen Planopilaris Among the Iranian Population

Maryamsadat Nejadghaderi1, Ashkan Tashk2 and Parvin Mansouri3, 1&3Tehran University of Medical Sciences (TUMS), Tehran, Iran, 2Mærsk McKinney Møller Institute (MMMI), University of Southern Denmark (SDU), Odense, Denmark

ABSTRACT

Introduction: The demographic of Lichen PlanoPilaris (LPP) among the Iranian population is unknown. The aim of this study is to describe the clinical, demographic, and histopathologic findings of lichen planopilaris in the Iranian population. Method: In this cross-sectional study, all the patients with Lichen planopilaris were referred to the dermatology clinic of Imam Khomeini hospital from 2013 to 2015. Their demographic characteristics, drug histories, onset of disease, and family histories were obtained by written questionnaire. Additionally, this study employed SPSS v.20 as the statistical analysis software. Results: One hundred patients were enrolled in this study. With an average age of 47.11 years, 78% of the patients were female, and 50 of these were housewives. The patients included were often from Tehran with Fars ethnicity. Among these patients, 7 had alopecia areata skin disease, and 10 of them suffered from thyroid disease. Most of the histopathology samples collected from these biopsies revealed degeneration of the basal layer of the follicular structure, perifollicular fibrosis, inflammatory cells, and atrophy of the pilosebaceous structures. Conclusion: Both the age spectrum and the disease distribution of LPP among the Iranian population were very diverse when compared to previous studies.

KEYWORDS

Lichen PlanoPilaris (LPP), Epidemiologic, Demographics, Clinical Features, histology


Adaptive Data Replication Optimization Based On Reinforcement Learning

Chee Keong Wee and Richi Nayak, Science & Engineering Faculty, Queensland University of Technology, Brisbane, Queensland, Australia

ABSTRACT

Data replication plays an important in enterprise IT landscapes where data is shared among multiple IT systems. IT administrators need to tune the replicating software’s configuration settingfor it to perform at its optimum level. it is also a bigger challenge to keep optimizing the software’ configuration constantly in order to keep up with the fluctuating workload in a dynamic business environment. we propose a novel approach of using reinforcement learning with meta-heuristics to create an adaptive optimization method for data replication software. The results show the replicating software that is managed by our proposed system can perform at an optimum level consistently under changing workloads throughout the period.

KEYWORDS

Data replication, optimization, meta-heuristics, deep reinforcement learning


Skin Cancer detection Using A hybrid deep Learning Approach: Combining convolutional neural Networks (CNN) And support vector machines (SVM)

HamitTaner Ünal, Adem Alpaslan Altun and Fatih Basçiftçi, Selcuk University, Konya, Turkey

ABSTRACT

Skin cancer is one of the most prevalent cancers in the world and requires dermoscopy-based visual diagnosis with high accuracy. Due to significant visual similarity between malignant skin tumors and benign skin lesions, it is very difficult to identify and differentiate between each other. In this work, we adopt a deep Convolutional Neural Network(CNN) integrated with a Support Vector Machine (SVM) approachto recognize skin cancer based on digital images.We use HAM10000 dataset for training the model and build a deep learning architecture where the last fully connected layer of CNN was replaced by a Support Vector Machine classifier to predict tumor types of the input images.In this model, CNN is used as a feature extractor unit and SVM performing as a recognizer tool.With this approach we aim to provide better diagnosis across different skin cancer categorieswhile providing better accuracy compared to traditional machine learning approaches.

KEYWORDS

deep learning, convolutional neural networks, support vector machines, skin cancer detection, artificial intelligence


Tibetan and Chinese Text Image Classification Based on Convolutional Neural Network

Jincheng Li1, Penghai Zhao1, Yusheng Hao2, Qiang Lin2, Weilan Wang1*, 1Key Laboratory of China's Ethnic Languages and Information Technology (Northwest Minzu University), Ministry of Education Lanzhou, P. R. China and 2College of Mathematics and Computer Science, Northwest Minzu University, Lanzhou, P. R. China

ABSTRACT

The first stage of Tibetan-Chinese bilingual scene text detection and recognition is the detection of TibetanChinese bilingual scene text. The detection results are mainly divided into three categories: successfully detected regions of Tibetan text and Chinese text, non-words regions with failed predictions. If the detected text image results are accurately classified, then the non-text images should be filtered in the recognition phase, meanwhile the Tibetan and Chinese text images can be identified by using different classifiers, such procedure can reduce the complexity of classification and recognition of two different characters by one recognition model. An accurate classification of Tibetan and Chinese text images is mattered. Therefore,this paper conducts a research on the classification of Tibetan, Chinese and non-text images by using convolutional neural networks. We perform a series of exploration about the classification accuracy of Tibetan, Chinese text images and non-text images with convolutional neural networks in different depths, and compare the accuracy with the classification results based on the transfer learning then analyze it. The results show that for the classification of Tibetan, Chinese and non-text images in the scene, using 7-layer convolutional neural network has reached saturation, and increasing the network depth does not improve the results, which provides reference values for Tibetan-Chinese text image classification.

KEYWORDS

Convolutional Neural Network, Tibetan-Chinese scene text image, image classification, transfer learning


Use of Blockchainin Public Key Infrastructure (PKI): A Systematic Literature Review

Nouf S. Aldahwan1 and DaniyalAlghazzawi2, 1Department of Information Systems, Faculty of Computing and Information Technology at King Khaled University, Abha, Saudi Arabia and 2Department of Information Systems, Faculty of Computing and Information Technology at King Abdulaziz University, Jeddah, Saudi Arabia

ABSTRACT

Blockchain technology has revolutionized the way people conduct transactions online. The distributed ledger technology has enabled the recording and tracking of resources and information even without a trustworthy authority as a central figure. Users can exchange transactions that are grouped into blocks following a particular sequence. The distributed append-only ledger allows messages to be recorded without reversal making it one of the most efficient technologies to share critical information and transact resources. An additional technology that has grown to become one of the most preferred security solutions, particularly for e-commerce transactions is Public Key Infrastructure (PKI). PKIs are complex systems comprising of multiple components that require coordination and integration of an organization’s business models. PKI uses two digital keys, one public, and another secret to ensure the confidentiality of a transaction, encryption, authentication, and signing of electronic data digitally. In this paper, we illustrated issues related to the PKI field, and then discussed how Blockchain technologies can fix this problem. We also illustrated the problems that occur with Blockchain being implemented.

KEYWORDS

Blockchain, Public Key Infrastructure (PKI)


Blockchains for DDoS Attacks Prevention in IoT: Systematic Literature Review

Maha Saad Alqhtani1 and Daniyal Alghazzawi2, 1Department of Information Systems, King Khalid University, Abha, Saudi Arabia and 2Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, Saudi Arabia

ABSTRACT

The rapid growth in the number of insecure portable and stationary IoT devices and a large increase in Internet traffic. The centralized approaches are not the most effective methods to prevent modern Distributed Denial of Service(DDoS) because it generates a large volume of traffic. This makes distributed denial of service a top security threat. An alternative to reduce the burden of detection and mitigation is to combine centralized defense systems, creating, cooperative and global protection system. However, existing approaches suffer from the complexity of deployment and operation across different systems. Blockchains appear in this scenario as an alternative used for the mitigation of DDoS attacks as it allows for the sharing of attack information in a fully a cooperative defense. In this paper, we present a comprehensive systematic literature review and measurement analysis of the Blockchain and DDoS protection in IoT.

KEYWORDS

Blockchains, Distributed Denial of Service, Internet of Thing.


Multi-label Classifier Performance Evaluation with Confusion Matrix

Damir Krstinic, Maja Braovic, Ljiljana Šeric and Dunja Božic-Štulic, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, R. Boškovica 32, Split 21000, Croatia

ABSTRACT

Confusion matrix is a useful and comprehensive presentation of the classifier performance. It is commonly used in the evaluation of multi-class, single-label classification models, where each data instance can belong to just one class at any given point in time. However, the real world is rarely unambiguous and hard classification of data instance to a single class, i.e. defining its properties with single distinctive feature, is not always possible. For example, an image can contain multiple objects and regions which makes multi-class classification inappropriate to describe its content. Proposed solutions to this set of problems are based on multi-label classification model where each data instance is assigned one or more labels describing its features. While most of the evaluation measures used to evaluate singlelabel classifier can be adapted to a multi-label classification model, presentation and evaluation of the obtained results using standard confusion matrices cannot be expanded to this case. In this paper we propose a novel method for the computation of a confusion matrix for multi-label classification. The proposed algorithm overcomes the limitations of the existing approaches in modeling relations between the classifier output and the Ground Truth (i.e. hand-labeled) classification, and due to its versatility can be used in many different research fields.

KEYWORDS

Classification, multi label classifier, performance evaluation, confusion matrix


Robust and Sparse Support Vector Machines via Mixed Integer Programming

Mahdi JAMMAL, Stephane CANU, Maher ABDALLAH

ABSTRACT

In machine learning problems in general, and in classification in particular, overfitting and inaccuracies can be obtained because of the presence of spurious features and outliers. Unfortunately, this is a frequent situation when dealing with real data. To handle outliers proneness and achieve variable selection, we propose a robust method performing the outright rejection of discordant observations together with the selection of relevant variables. A natural way to define the corresponding optimization problem is to use the `0 norm and recast it as a mixed integer optimization problem (MIO) having a unique global solution, benefiting from algorithmic advances in integer optimization combined with hardware improvements. We also present an empirical comparison between the `0 norm approach, the 0-1 loss and the hinge loss classification problems. Results on both synthetic and real data sets showed that, the proposed approach provides high quality solutions.

KEYWORDS

Robust Classification, Sparse Classification, SVM, Mixed Integer Programming


Effective Teaching Startegies in an Online Enviroment

Dr. Keith Buckley, Director of Physical Education Rollins College, USA

ABSTRACT

The online classroom is a powerful teaching and learning arena in which new practices and new relationships can make significant contributions to learning. Instructors must be trained not only to use technology, but also to shift the ways in which they organize and deliver material. Making this shift can increase the potential for learners to take charge of their own learning process and facilitate the development of a sense of community. In constructing an online course, the instructor must take into account: designing the course, implementing content, facilitating learning, relevant assignments and course evaluation. A learner-centered approach acknowledges what students bring to the online classroom, their background, needs, and interests, and what they take away as relevant and meaningful outcomes. This paper will explore how to design implement and instruct a thorough online course that engages the student and gives them a platform to learn and enhance their academic experience.

KEYWORDS

Online Learning, Education, Teaching


Fusion Multi Focus Images With Neighbor Local Distance

Ias Sri Wahyuni1 and Rachid Sabre2, 1Universitas Gunadarma, Jl. Margonda Raya No. 100 Depok 16424, Indonesia and 2Laboratory Béogéosciences CNRS, University of Burgundy/Agrosup Dijon, France

ABSTRACT

The aim of multi-focus image fusion is to integrate images with different objects in focus so that obtained a single image with all objects in focus. In this work, we present a novel multi-focus image fusion method based on neighbor local variability (NLV). This method takes into consideration the information in the surrounding region of pixels. Indeed, at each pixel, the method exploits the local variability calculated from quadratic difference between the value of pixel and the value of all pixels that belong to its neighborhood. It expresses the behavior of pixel relative to all pixels belong to its neighborhood. The variability preserves edge feature because it detects the abrupt image intensity. The fusion of each pixel is performed by weighting each pixel by the exponential of the local variability. The precision of this fusion is depending on the large of the neighborhood where the large depends on the blurring characterized by the variance and its size of blurring filter. We construct a model that give the value of the large from the variance and the size of blurring filter. We compare our method with other methods, we show that our method gives the best result.

KEYWORDS

Neighbor Local Variability, Multi-focus image fusion, RMSE


Simple Face Thermal Features for Gender Discrimination

Georgia Koukiou and Vassilis Anastassopoulos, Electronics Laboratory Physics Department, University of Patras, Greece

ABSTRACT

A very simple approach is proposed for gender discrimination using thermal infrared images of the persons' face. The selected features are actually based on the mean value of the pixels of specific locations on the face. It is proved that the discrimination is feasible either by simple visualization in the feature space or by using a relatively simple neural network for this purpose. All cases of persons from the used database, males and females, are correctly distinguished based on the mean value of the employed locations of the face. Classification results are verified using two conventional approaches, namely: a. the simplest possible neural network so that generalization is achieved along with successful discrimination between all persons and b. the leave-one-out approach to demonstrate the classification performance on unknown persons using the simplest classifiers possible.

KEYWORDS

Thermal Infrared, Face Recognition, Gender Discrimination, Neural Networks


Regular Plans with Differentiated Services using Cuckoo Algorithm

John Tsiligaridis, Department of Computer Science, Heritage University, Toppenish, WA, USA

ABSTRACT

The management of a server’s differentiated services is one of the top interesting roles in mobile technology. A server with cluster resources serves clients in groups of multiple classes according to their service level agreement. The server will have the ability to create a Regular Broadcasting Plan (RBP) after some data transformation for either a single or multiple channels. Based on theorems the Cuckoo Search (CS) algorithm discovers RBPs with delay differentiated services (CSDS). Given the delay ratio of just two services, the Extended version of CSDS (ECSDS), can discover the RBP with the smallest waiting time for low priority mobile services. These algorithms will enhance servers’ ability for self-monitoring, and self-organizing adapting their services to predefined ratio parameters.

KEYWORDS

Broadcasting, Cuckoo Algorithm, Differentiated Services , Mobile Computing, Algorithms


Countermeasures using Continuous Time Markov Chains

Karim Lounis, Queen’s University, Canada

ABSTRACT

ADTrees (Attack-Defense Trees) are graphical security mod-eling tools used to logically represent attack scenarios along with theircorresponding countermeasures in a user-friendly way.Many researchersnowadays use ADTrees to represent attack scenarios and perform quan-titative as well as qualitative security assessment. Among all different existing quantitative security assessment techniques, CTMC (Continu-ous Time Markov Chains) has been attractively adopted for ADTrees.ADTrees are usually transformed into CTMCs, where traditional stochas-tic quantitative analysis approaches can be applied. For that end, thecorrect transformation of an ADTree to a CTMC requires that each in-dividual element of an ADTree should have its correct and complete representation in the corresponding CTMC. In this paper, we mainlyfocus on modeling counter measures in ADTrees using CTMCs. The ex-isting CTMC-model does not provide a precise and complete model-ing capability, in particular, when cascaded-countermeasures are used.Cascaded-counter measures occur when an attacker and a defender in a given ADTree recursively counter each other more than one time in a given branch of the tree. We propose the notion of tokenized-CTMC to construct a new CTMC-model that can precisely model and represent countermeasures in ADTrees. This new CTMC-model allows to handle cascaded-countermeasure scenarios in a more comprehensive way.

KEYWORDS

Attack-defense trees, CTMC, security graphical models, stochas-tic models, and quantitative security assessment.