SEREF SAGIROGLU | Gazi University (original) (raw)

Papers by SEREF SAGIROGLU

2017 5th International Symposium on Digital Forensic and Security (ISDFS), 2017

Data is the most valuable asset for organization and needs to be secured. Especially organization... more Data is the most valuable asset for organization and needs to be secured. Especially organizations should pay attention to data leakage issues because the leakage of sensitive data could give great damages or harm more than expected. There have been various factors causing data leakage but the recent research showed that the rate of insider attacks was the highest score and most of the data leakage incidents occurred in network based channels such as e-mails, clouds, instant messaging, etc. Insiders can consider transform or modify sensitive data. The main purpose of these attacks is to leak sensitive data from the systems without any alert of security systems. This paper aims to detect modification attacks on sensitive words, solve the data leakage problem and proposes a cascaded Data Leakage Prevention (DLP) system for Turkish language. The system consists of training and detecting phases. In training phase, a list of sensitive words was generated from the sensitive document sets. In detection phase, the main purpose is to detect modified sensitive content that attacker aims to bypass security systems. The types of possible attacks such as adding, deleting and changing characters in the sensitive words, deleting whitespaces from both sides of sensitive words and adding whitespace to the middle of the sensitive words were considered for designing the system. Boyer Moore (BM) algorithm was used to search exact sensitive strings exposed to whitespace attack and Smith Waterman (SW) sequential alignment algorithm was also employed to detect modified string attacks. TF-IDF method was used to extract the sensitive words of sensitive documents. Latent Semantic Indexing (LSI) was preferred to model document topics. For extracting and analyzing Turkish, Zemberek was used for. The results have shown that the proposed DLP system supporting Turkish language has a plausible solution.

Turkish Neurosurgery, 2020

AIM During daily practice of neurosurgery, especially in sub-specializations such as radiosurgery... more AIM During daily practice of neurosurgery, especially in sub-specializations such as radiosurgery, tumor segmentation is essential and time consuming. Artificial intelligence methods are getting widely used for such purposes. In this study we propose a deep convolutional generative adversarial networks (DCGAN) model which learns normal brain MRI from normal subjects than finds distortions such as a glioma from a test subject while performing a segmentation at the same time. MATERIAL AND METHODS MRIs of 300 healthy subjects were employed as training set. Additionally, test data were consisting anonymized T2-weigted MRIs of 27 healthy subjects and 27 HGG patients. Consecutive axial T2-weigted MRI slices of every subject were extracted and resized to 364x448 pixel resolution. The generative model produced random normal synthetic images and used these images for calculating residual loss to measure visual similarity between input MRIs and generated MRIs. RESULTS The model correctly detected anomalies on 24 of 27 HGG patients' MRIs and marked them as abnormal. Besides, 25 of 27 healthy subjects' MRIs in the test dataset detected correctly as healthy MRI. The accuracy, precision, recall, and AUC were 0.907, 0.892, 0.923, and 0.907, respectively. CONCLUSION Our proposed model demonstrates acceptable results can be achieved only by training with normal subject MRIs via using DCGAN model. This model is unique because it learns only from normal MRIs and it is able to find any abnormality which is different than the normal pattern.

2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA), 2016

Biometric systems may be used to create a remote access model on devices, ensure personal data pr... more Biometric systems may be used to create a remote access model on devices, ensure personal data protection, personalize and facilitate the access security. Biometric systems are generally used to increase the security level in addition to the previous authentication methods and they seen as a good solution. Biometry occupies an important place between the areas of daily life of the machine learning. In this study;the techniques, methods, technologies used in biometric systems are researched, machine learning techniques used biometric aplications are investigated for the security perspective, the advantages and disadvantages that these tecniques provide are given. The studies in the literature between 2010-2016 years, used algorithms, technologies, metrics, usage areas, the machine learning techniques used for different biometric systems such as face, palm prints, iris, voice, fingerprint recognition are researched and the studies made are evaluated. The level of security provided by the use of biometric systems by developed using machine learning and disadvantages that arise in the use of these systems are stated in detail in the study. Also, impact on people of biometric methods in terms of ease of use, security and usages areas are examined.

DergiPark (Istanbul University), 2008

Literatürde biyometrik tanıma sistemlerine yönelik birçok başarılı teknik, yaklaşım ve algoritma ... more Literatürde biyometrik tanıma sistemlerine yönelik birçok başarılı teknik, yaklaşım ve algoritma geliştirilmiş ve gerçekleştirilmiştir. Bu çalışmalar incelendiğinde biyometrik özellikler arasındaki ilişkinin analizine yönelik herhangi bir çalışmaya rastlanmamaktadır. Sunulan çalışmada parmakizi, yüz, iris, retina ve el geometrisi gibi biyometrik özellikler arasında olabilecek herhangi bir ilişkinin varlığı tartışılmakta ve kişilerin yalnızca parmakizini kullanarak yüzlerini tahmin etmeye yönelik yapay sinir ağları temelli yeni ve zeki bir sistem tanıtılmaktadır. Elde edilen sonuçlar, sunulan çalışmanın bu konuda gerçekleştirilmiş bir ilk çalışma olmasına ve sonuçların hiç bir son işlemden geçirilmeden en ham şekliyle sunulmasına rağmen başarısının kabul edilebilir niteliklerde olduğunu ve gelecekte konuyla ilgili farklı çalışmaların geliştirilmesine katkı sağlayacağını göstermektedir.

International Journal of Information Technology & Decision Making, 2016

Modeling user knowledge and creating user profiles not only for special web-based social media bu... more Modeling user knowledge and creating user profiles not only for special web-based social media but also for complex and mixed personalized learning environments are important research challenges. The key component for adaptation is the user’s knowledge model. This paper introduces fuzzy metric (FM)-based novel and efficient similarity measurement method and adaptive artificial neural network (AANN) and artificial bee colony (ABC)-based knowledge classification approaches for personalized learning environments. For this purpose, FM-based method has been developed to measure distances more efficiently among the users and their knowledge model using the web logs/session data. In addition, a novel knowledge classifier based on ABC and AANN having combined with the generic object model has been developed for user modeling strategies and user modeling server of adaptive educational electric course (AEEC). Finally, the approaches have been tested to compare the classification performance o...

Applied Energy, 2015

The attention on the smart grids and smart grid technologies has grown significantly over the las... more The attention on the smart grids and smart grid technologies has grown significantly over the last few years. The analysis made in this study is grounded on the smart grid projects database of the Joint Research Centre (European Commission). The European smart grid projects are analyzed among others in terms of: number, countries, duration and collaboration. Additionally, an analysis is done regarding the annual number of starting and concluded/planned to be concluded projects, the total number of participants per year, the distribution of smart grid applications per stage of development, year and EU country and an overview of the investments in the European smart grid projects is provided. Afterwards a forecast is done regarding the number of projects. As a result of graphical and predictive analyses, many essential inferences are achieved related to the current status and the anticipated short-term trends of smart grid projects.

2009 3rd IEEE International Conference on E-Learning in Industrial Electronics (ICELIE), 2009

This study introduces to usage of an up-to-date software platform named .NET for the Web-based le... more This study introduces to usage of an up-to-date software platform named .NET for the Web-based learning models directed to education of a more special area that includes electrical machines courses in where dynamic and interactive simulations are also included. Users can perform several didactic educational activities by means of these dynamic simulations. To illustrate the system, an interactive dc motor

As many approaches and algorithms for biometric recognition techniques have been developed and pr... more As many approaches and algorithms for biometric recognition techniques have been developed and proposed in the literature in details, the relationships among biometric features have not been studied yet. This study presents an analysis of existence of any relationship among biometric features like fingerprints and faces. A new and novel approach based on artificial neural networks was designed and introduced to generate faces from fingerprint images. Experimental results have shown that faces were recognised with high accuracy from only fingerprints. Although the proposed system was an initial work on this topic, the results were very encouraging and promising for the future studies even if any post-processing was not applied.

2014 13th International Conference on Machine Learning and Applications, 2014

In this study an approach that uses social networking data for developing sentiment analysis syst... more In this study an approach that uses social networking data for developing sentiment analysis system is proposed. With the help of developed software, it is tried to find out whether there is any relation between universities' academic success and sentiment of the public about them in social media. After collecting enough text based data from Twitter, preprocessing of data is carried out and final data is trained by means of Naïve Bayes Classifier. After testing process, experimental results have shown that developed sentiment analysis system can classify the tweets about top 10 universities according to URAP rankings in terms of their sentiment with the 72.33% success rate, and proposed methodology can be used by universities for understanding sentiment of the public about them in social media.

2008 International Symposium on Power Electronics, Electrical Drives, Automation and Motion, 2008

This study introduces original system architecture for the web based remote experimentation appli... more This study introduces original system architecture for the web based remote experimentation applications for e-learning systems. The speed control of a traveling wave ultrasonic motor (USM) is achieved as a remote experimental study over the Internet to illustrate the architecture. For this reason, a microcontroller based digitally controlled drive system for the USM is designed firstly. A data acquisition board is then used to perform data transferring and control operations between the server and the drive system. Data obtained from experimental set is analyzed in the MATLAB program installed on the server. Similarly, MATLAB web server (MWS) software is used to send and receive data over the Internet. Since all operation steps are executed from the server, clients do not need any special hardware or software unit. Thus, the system developed has an independent structure for the clients. Remotely located clients can operate and control the motor and conduct the experiment. Additionally, real laboratory environment and the experimental set can be monitored on the client computer's screen while the experiment is being operated. The results obtained from the tests show that the system developed has significant features such as simplicity, sensitivity, reliability and speedy. Accordingly, the system presented can be adapted to elearning systems easily as an effective tool for remote experimentation. Index Terms-Remote experiment, virtual laboratory, real time system, ultrasonic motor, speed control.

Lecture Notes in Computer Science

Abstract. Even if biometric features have been deeply studied, tested and successfully applied to... more Abstract. Even if biometric features have been deeply studied, tested and successfully applied to many applications, there is no study in achieving a biometric feature one from another. This study presents a novel intelligent approach analysing the existence of any relationship ...

2008 IEEE 16th Signal Processing, Communication and Applications Conference, 2008

... Parmakizi, yüz, kulak, iris, retina, el geometrisi, ses gibi teknikleri kapsayan biyometri, k... more ... Parmakizi, yüz, kulak, iris, retina, el geometrisi, ses gibi teknikleri kapsayan biyometri, kişilerin fiziksel veya davranışsal özellikleri kullanılarak kimliklendirilmesi bilimi olarak açıklanabilir [1 ... [16] Sağıroğlu, Ş., Beşdok, E. ve Erler, M., Mühendislikte Yapay Zeka Uygulamaları I ...

2012 11th International Conference on Machine Learning and Applications, 2012

ABSTRACT Up to now a large number of approaches and algorithms about biometric recognition techni... more ABSTRACT Up to now a large number of approaches and algorithms about biometric recognition techniques have been developed and proposed in the literature in details. Relationships about biometrics have recently been studied in the field by author research group. This paper presents a web based intelligent face mask prediction system based on artificial neural networks to predict faces from only fingerprint images. The aim of this study is to develop an online platform to be used by users, researchers or academics around the world. The results have been shown that the system can be used for live test through internet. The proposed web based system is successful to support and encourage researchers around the world to see and exercise this new concept in biometrics. The developed system predicts an appropriate face mask covering nose, eyes, ears and mouth from fingerprint images with the help of an intelligent model.

Sixth International Conference on Machine Learning and Applications (ICMLA 2007), 2007

In this study, a web based maze robot system has been designed and implemented for solving differ... more In this study, a web based maze robot system has been designed and implemented for solving different maze algorithms with the help of machine learning approaches. The robot system has a map-based heuristic maze solving algorithm. The algorithm used for solving the maze is ...

2006 5th International Conference on Machine Learning and Applications (ICMLA'06), 2006

This work presents an intelligent automatic fingerprint identification and verification system ba... more This work presents an intelligent automatic fingerprint identification and verification system based on artificial neural networks. In this work, the design processes of the system have been presented step by step. In order to make the system automatic, software was developed for ...

2012 International Conference on Renewable Energy Research and Applications (ICRERA), 2012

ABSTRACT Wind speed forecasting is required for ensuring an efficient utilization of the wind pow... more ABSTRACT Wind speed forecasting is required for ensuring an efficient utilization of the wind power generated by wind turbines. This paper purposes the short-term wind speed forecasting in a 2-dimesional input space using the developed k-nearest neighbor (k-NN) classifier. As well, the effects of the nearest neighbor number and the selected distance metric on the wind speed forecasting were analyzed and many useful inferences were mined in order to minimize the forecasting error. The results have shown that the k-NN classifier which uses wind direction and relative humidity parameters achieved the best forecasting results for k=10 in the Minkowski distance metric. On the other hand, the k-NN classifier which uses wind direction and atmosphere pressure parameters gave the worst forecasting results for k=1 in the Euclidean distance metric.

Robotica, 1995

SummaryThis paper discusses three methods of training multi-layer perceptrons (MLPs) to model a s... more SummaryThis paper discusses three methods of training multi-layer perceptrons (MLPs) to model a six-degrees-of- freedom inertial sensor. Such a sensor is designed for use with a robot to determine the location of objects it has to pick up. The sensor operates by measuring parameters related to the inertia of an object and computing its location from those parameters. MLP models are employed for part of the computation. They are trained to output the orientation of the object in response to an input pattern that includes the period of natural vibration of the sensor on which the object rests. After reviewing the working principle of the sensor, the paper describes the three MLP training methods (backpropagation, optimisation using the Levenberg-Marquardt algorithm, evolution based on the genetic algorithm) and presents the experimental results obtained.

International Journal of Machine Tools and Manufacture, 2001

This paper presents an overview of four algorithms used for training multilayered perceptron (MLP... more This paper presents an overview of four algorithms used for training multilayered perceptron (MLP) neural networks and the results of applying those algorithms to teach different MLPs to recognise control chart patterns and classify wood veneer defects. The algorithms studied are Backpropagation (BP), Quickprop (QP), Delta-Bar-Delta (DBD) and Extended-Delta-Bar-Delta (EDBD). The results show that, overall, BP was the best algorithm

Expert Systems with Applications, 2009

This paper introduces a new vision-based and web-based mobile robot platform. The platform consis... more This paper introduces a new vision-based and web-based mobile robot platform. The platform consists of control and communication centers, a mobile robot and real-time support libraries. All activities in the platform are achieved by only computer vision techniques. The platform provides monitoring, tele-controlling and programming for real-time educational exercises and helps to the users to achieve these exercises through a standard web browser without any need for additional support software. The results have shown that the proposed, designed and implemented platform provide amazing new facilities and features to the users (students and researchers) in applying their real-time exercises on web.