Harun Artuner

Active authentication by mouse movements

We present an application for authenticating users with mouse movements. We extract the features from nine paths between seven squares displayed consecutively on the screen. Although the paths among the squares are fixed, user can not get familiar to the paths by the selection algorithm of next path. This prevents similar results from different users. We experimented on improving the success rate, by selecting the optimal time interval between mouse movement points and refining features to obtain valid data. Experiments results presented a false acceptance rate (FAR) of 5.9% and a false reject rate (FRR) of 5.9%.

Clustering of volcanic ash arising from different fragmentation mechanisms using Kohonen self-organizing maps

In this study, we present the visualization and clustering capabilities of self-organizing maps (SOM) for analyzing high-dimensional data. We used SOM because they implement an orderly mapping of a high-dimensional distribution onto a regular low-dimensional grid. We used surface texture parameters of volcanic ash that arose from different fragmentation mechanisms as input data. We found that SOM cluster 13-dimensional data more accurately than conventional statistical classifiers. The component planes constructed by SOM are more successful than statistical tests in determining the distinctive parameters.

Effects of Current and Potential Dental Etchants on Nerve Compound Action Potentials

In this study, a 35% phosphoric acid gel (3M Scotchbond etchant), a nonrinse etchant (NRC), and two EDTA-containing conditioners (RC-Prep and File-Eze) were tested in vitro for blocking nerve conductance evoked in the rat sciatic nerve after local application. The phosphoric acid gel and NRC completely and irreversibly inhibited conductance. On the other hand, RC-Prep reduced the compound action potentials (cAPs) by 50% in 120 min. With File-Eze, the reduction in cAPs was less than 50% after an application time of 160 min (61.8 +/- 1.8%). At 160 min the cAPs in the RC-Prep group had been inhibited by 62.4%. These results indicated strong neurotoxic effects of phosphoric acid and NRC when applied directly on exposed pulp in the total etch procedure.

Application of Decision Tree Algorithm for classification and identification of natural minerals using SEM-EDS

A mineral is a natural, homogeneous solid with a definite chemical composition and a highly ordered atomic arrangement. Recently, fast and accurate mineral identification/classification became a necessity. Energy Dispersive X-ray Spectrometers integrated with Scanning Electron Microscopes (SEM) are used to obtain rapid and reliable elemental analysis or chemical characterization of a solid. However, mineral identification is challenging since there is wide range of spectral dataset for natural minerals. The more mineralogical data acquired, time required for classification procedures increases. Moreover, applied instrumental conditions on a SEM-EDS differ for various applications, affecting the produced X-ray patterns even for the same mineral. This study aims to test whether C5.0 Decision Tree is a rapid and reliable method algorithm for classification and identification of various natural magmatic minerals.Ten distinct mineral groups (olivine, orthopyroxene, clinopyroxene, apatite, amphibole, plagioclase, K-feldspar, zircon, magnetite, biotite) from different igneous rocks have been analyzed on SEM-EDS. 4601 elemental X-ray intensity data have been collected under various instrumental conditions. 2400 elemental data have been used to train and the remaining 2201 data have been tested to identify the minerals. The vast majority of the test data have been classified accurately. Additionally, high accuracy has been reached on the minerals with similar chemical composition, such as olivine ((Mg,Fe)2SiO4]) and orthopyroxene ((Mg,Fe)2SiO6]). Furthermore, two members from amphibole group (magnesiohastingsite, tschermakite) and two from clinopyroxene group (diopside, hedenbergite) have been accurately identified by the Decision Tree Algorithm. These results demonstrate that C5.0 Decision Tree Algorithm is an efficient method for mineral group classification and the identification of mineral members. C5.0 Algorithm is tested as a method for mineral identification using EDS data.Selected minerals have been classified accurately using C5.0 Algorithm.Effects of instrumental conditions have been minimized by applied methods.C5.0 Decision Tree stands as an effective tool for mineral classification using EDS.

A New Parallel Matrix Multiplication Algorithm for Wormhole-Routed All-Port 2D/3D Torus Networks

A new matrix multiplication algorithm is proposed for massively parallel supercomputers with 2D/3D, all-port torus interconnection networks. The proposed algorithm is based on the traditional row-by-column multiplication matrix product model and employs a special routing pattern for better scalability. It compares favorably to the variants of Cannon’s and DNS algorithms since it allows matrices of the same size to be multiplied on a higher number of processors due to lower data communications overhead.

Space weather studies of IONOLAB group

IONOLAB is an interdisciplinary research group dedicated for handling the challenges of near earth environment on communication, positioning and remote sensing systems. IONOLAB group contributes to the space weather studies by developing state-of-the-art analysis and imaging techniques. On the website of IONOLAB group, www.ionolab.org, four unique space weather services, namely, IONOLAB-TEC, IRI-PLAS-2015, IRI-PLAS-MAP and IRI-PLAS-STEC, are provided in a user friendly graphical interface unit. Newly developed algorithm for ionospheric tomography, IONOLAB-CIT, provides not only 3-D electron density but also tracking of ionospheric state with high reliability and fidelity. The algorithm for ray tracing through ionosphere, IONOLAB-RAY, provides a simulation environment in all communication bands. The background ionosphere is generated in voxels where IRI-Plas electron density is used to obtain refractive index. One unique feature is the possible update of ionospheric state by insertion of Total Electron Content (TEC) values into IRI-Plas. Both ordinary and extraordinary paths can be traced with high ray and low ray scenarios for any desired date, time and transmitter location. 2-D regional interpolation and mapping algorithm, IONOLAB-MAP, is another tool of IONOLAB group where automatic TEC maps with Kriging algorithm are generated from GPS network with high spatio-temporal resolution. IONOLAB group continues its studies in all aspects of ionospheric and plasmaspheric signal propagation, imaging and mapping.

Recognition of radio signals with deep learning Neural Networks

Nowadays, the use of Artificial Neural Networks shows continuity in many different applications. In particular, the use of deep learning type neural networks, seen as impractical previously, is increasing by processing power growth. Software Defined Radio (SDR) is another hot topic called as Digital Radio concept. With a growing interest in the environment was formed especially by radio amateurs. Analyzing signals and making sense out of them by SDR is accepted in many areas. In this study, design and implementation of a sample work has been subjected for identification of received radio signals on SDR with deep learning neural networks. (DLNN).

A syllable-based Turkish speech recognition system by using time delay neural networks (TDNNs)

In this paper, we present a model for Turkish speech recognition. The model is syllable-based, where the recognition is performed through syllables as speech recognition units. The main goal of the model is to recognize as much as possible of a given continuous speech by identifying only a small set of syllables in the language. For that purpose, only the syllable types with a higher frequency are selected for the recognition. The use of longer recognition units in speech recognition systems increases the success of the recognition since it is easier to detect the endpoints of syllables when compared to phonemes. On the other side, word-based recognition requires a very large dataset that includes all the words and word forms in the language, which is also another challenge. Hereby, we take the advantage of Turkish being an ortographically transparent and syllabified language. Our model employs time delay neural networks (TDNNs) for learning syllables. We achieve an accuracy of %65.6 on our large vocabulary continuous speech corpus. In addition, we define an algorithm for the automatic detection of syllable boundaries which gives an accuracy of %44. The automatic syllable boundary detection module is used for the recognition of isolated syllables rather than a continuous speech.

Efficient mass-balancing of flowsheet data

In this paper, we discuss the general formulation and solution of linear mass balance equations for mass-separators. One-input/two-output and two-input/two-output models with multi-component flows are discussed and a simple, yet efficient solution method is proposed. The proposed method does not require a separate gross error detection mechanism or any statistical knowledge derived from historical plant data. Numerical examples are presented with relevant MATLAB code fragments. Obtained results compare favorably to those provided by JKSimMet (Version 5.1) software.

Dağıtılmış Anahtarlamalı Özdirenç Görüntüleme Sistemi

Within last three decades, direct current resistivity imaging method has been subjected to a considerable development both due to the instrumental evolution and due to the increase in the speed and the capacity of the computation. The application is generally made using a central switching system or another switching system which is distributed over electrodes, called smart electrode system. In terms of field surveys, both systems have advantages and disadvantages. Smart electrode systems may be advantageous due to the possibility of producing lightweight cables. The smart electrode system developed in the Hacettepe University is 75% lighter than the equivalent systems that are operating with a central switching unit. Principally, system is composed of a small central control unit, smart electrode units that are connected to the electrodes and the main cable. Inside the main cable, there lie 6 inner cables; two of them are isolated due to interference. System runs automatically via a laptop computer which is serially connected to the control unit. The system was tested with short and long profiles and observed to be working properly. The lightweight of the system provides a considerable ease especially during the measurement İ. Ulusoy e-posta: inan@hacettepe.edu.tr