Event program

Tuesday, 5/22/2012 11:00 AM - 1:00 PM, Camelia 2, Hotel Adriatic, Opatija 11:00 AM - 1:00 PM Papers  1. J. Okroša (Filozofski fakultet, Zagreb, Croatia) Djeluje li laskanje? O utjecaju snage poruke i pripisivanja identiteta altruistične osobe na odaziv sudionika u istraživanju putem interneta  U reklamnim porukama česta je emocionalna nagrada obećanje željenog identiteta. Budućem korisniku može se učiniti privlačnim pripadanje skupini mladih urbanih vlasnika mobitela te mu se reklama tako i obraća: kao pripadniku opisane skupine. Cilj istraživanja je ispitati kako na odaziv u online istraživanju djeluje takvo pripisivanje identiteta u uvjetima slabih i jakih argumenata. U istraživanju je korišten pripisan identitet altruistične osobe. Hipoteza je će pripisan identitet djelovati u situaciji kada su u pozivu sudionicima na istraživanje putem interneta navedeni slabi argumenti o važnosti istraživanja i to tako da će odaziv istraživanju biti veći. U situaciji kada su u pozivu sudionicima na istraživanje putem interneta navedeni jaki argumenti neće doći do utjecaja pripisanog identiteta na povećanje odaziva. Odaziv istraživanju operacionaliziran je kao broj sudionika koji su upitniku pristupili te broj ispitanika koji su upitnik dovršili (od ispitanika koji su upitniku pristupili). Istraživanje je pokazalo da više studenata pristupa upitniku u slučaju kada poruka ne sadrži pripisani identitet, neovisno o kvaliteti sadržanih argumenata. Kod mjere dovršavanja upitnika pokazuje da u slučaju slabih argumenata više studenata dovršava upitnika kada je pripisan identiteta. U slučaju jakih argumenata više studenata dovršava upitnika kada identitet altruistične osobe nije pripisan. 2. T. Težak (Učiteljski fakultet, Zagreb, Croatia), M. Dumančić (Učiteljski fakultet u Zagrebu, Zagreb, Croatia) Razvoj modela e-škole kao dodatnog oblika nastave za djecu koja žive izvan Hrvatske  U radu je prikazan model e-škole koji bi prvenstveno koristio djeci koja žive izvan Republike Hrvatske kao dodatni oblik nastave, a daljnjim razvojem bi bio primjenjiv učenicima u slabije naseljenim područjima RH i učenicima koji su djelomično ili u potpunosti onemogućeni u praćenju redovne nastave. Prijenos nastavnog procesa temelji se na sustavu učenja na daljinu posredstvom multimedije, a u radu su opisani primjenjivi alati za e-učenje na daljinu, osnovni „servisi“ i sadržaji škole, oprema potrebna za rad škole i oprema potrebna za učenike škole. Cilj ovakve škole nije zamjena klasične škole i klasičnog oblika nastave nekom novom školom, već dopuna školskog sustava u područjima u kojima se redovnim putem nastava ili ne može održavati ili je otežana, a naglasak je na nastavi koja se održava u „realnom vremenu“ bez obzira na udaljenost učitelja i učenika, te na niskim troškovima opreme koja je potrebna za praćenje nastave. 3. L. Tepeš Golubić, J. Kolarec (TVZ, Zagreb, Croatia) Tehnički i socijalni pogledi na web forume, na primjeru foruma TVZ-a  Društveni proces koji se odvija u današnje doba izazvan je pojavom Internet komunikacije koja se odvija od komunikacije pojedinca sa pojedincem iz iste sredine, pa sve do komunikacije pojedinca sa grupom ljudi iz neke druge sredine. Virtualna komunikacija postupno formira virtualno društvo u kojemu se postavljaju neke nove društvene norme i novi „jezik“ kojime se koriste te svakim danom sve veće virtualne zajednice. Naša svakodnevna komunikacije sve se više referencira na tu virtualnu komunikaciju koja se događa na forumima. Jedan dio komunikacije odvija se na forumima kojima pružaju prividnu anonimnost i omogućuju postavljanje raznih pitanja točno određenoj skupini bez straha od odmazde sredine. Zajednice koje se kriju pod pojmom foruma moramo gledati kao uzročno posljedičnu vezu između društva i Interneta, one nam zapravo ne otkrivaju samo tehnološku stranu informatičke pismenosti našeg društva, nego otkriva koliko smo mi zapravo usamljeni/izgubljeni u velikom dotoku novih informatičkih tehnologija. Rad je rezultat istraživanja studenskog foruma Tehničkog veleučilišta Zagreb, forum.sztvz.hr. 4. M. Kreković, P. Čerić, T. Dominko, M. Ilijaš, K. Ivančić, V. Skolan, J. Šarlija (Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia, Zagreb, Croatia) A Method for Real-Time Detection of Human Fall from Video  A method for real-time detection of human fall from video Miranda Kreković, Paolo Čerić, Tena Dominko, Matija Ilijaš, Kristijan Ivančić, Viktor Skolan, Josip Šarlija E-mail: miranda.krekovic@fer.hr Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia Summary In this paper we present a method for real-time detection of human fall from video for support of elderly people living alone in their homes. The detection algorithm has four steps: background estimation, extraction of moving objects, object feature extraction, and fall detection. The detection is based on features that quantify dynamics of human motion and body orientation. Algorithms are implemented in C++ using the OpenCV library. The method is tested using a single camera and 18 test video recordings showing typical fall scenarios and regular household behaviour. The experimental results show 94% of human fall detection accuracy. 5. A. Divjak, M. Maričević (Grafički fakultet Sveučilišta u Zagrebu, Zagreb, Croatia) LightDesk - sustav za računalnu interakciju visoke rezolucije baziran na infracrvenom raspoznavanju oblika  LightDesk je sustav za prikaz slike visoke rezolucije sa mogućnošću digitalnog crtanja, slikanja i generalne računalne interakcije, jednostavnog dizajna i niske cijene. Temelji se na višestrukoj pozadinskoj digitalnoj projekciji i raspoznavanju oblika, te se u radu ponaša poput LCD grafičkog tableta. Interakcija se vrši pomoću mehaničke olovke napravljene pomoću 3D tiska sa integriranim infracrvenim osvjetljenjem, a raspoznavanje nagiba i pritiska olovke ostvaruje se preko raspoznavanja slike koju snima pet infracrvenih kamera ugrađenih u sustav. Uz radni prostor rezolucije 3840*1536 piksela i dijagonale 135 cm (53 inča), te prepoznavanje 512 razina pritiska i promjenu nagiba od 0.5 stupnja, sustav omogućava iznimno visoku razinu produktivnosti u primjenama poput 3D grafike, animacije i vizualizacije, CAD-a, CAE-a, te obrade slika i videa. Korištenjem inovativnih rješenja za prepoznavanje interakcije olovke sa radnom površinom, u kombinaciji sa ergonomskim dizajnom, stvoren je sustav koji je iznimno ugodan za korištenje, te pruža veliku slobodu i lakoću korištenja modernih računalnih aplikacija koje zahtijevaju preciznost pri izradi sadržaja. Ovaj rad prvenstveno demonstrira kako je moguće pomoću komercijalno dobavljivih komponenti i 3D tiska stvoriti sustav koji performansama i cijenom predstavlja iznimno primamljivu opciju za prikaz slike i interakciju sa računalnom. 6. M. Šošić, M. Šikić (FER, Zagreb, Croatia) CUDA implementation of the algorithm for simulating the epidemic spreading over large networks  Preventing and controlling of epidemics in human or computer networks is a top problem. This paper presents an implementation of the algorithm that simulates the epidemic spreading in networks, using CUDA architecture. Spreading of the epidemics over the network is modeled by a discrete SIR (Susceptible - Infected - Recovered) model. This implementation enables selection of starting nodes and monitoring the epidemic spread in each cycle. In comparison to a common CPU implementation, the CUDA implementation achieves about 10x faster execution time, which is of high significance to running tests on larger networks. The implementation was tested on real social networks consisting of more than 10 million nodes. We believe that this implementation can have practical value in analysis of the epidemic spreading over large networks. 7. N. Banić (Faculty of Electrical Engineering and Computing, Zagreb, Croatia) Detection of Commercials in Video Content Based on Logo Presence Without Its Prior Knowledge  Among many methods for commercial detection in TV video content the most precise and most used one in software today is the method of recognizing logo presence, and freeware as well as shareware shoftware using it usually needs some data about the channel logo. In this paper a commercial detection method is described which is also based on logo recognition, but without need of any prior knowledge of logo data like shape, position or colors. Applying the method on several test videos showed a 100% recognition of commercial presence and about 80% of correct determination of the begin and the end of commercials whereby videos with noise were also used. 8. I. Sluganović, A. Karlović, M. Šare, P. Bosilj, S. Horvat (Sveučilište u Zagrebu, Fakultet elektrotehnike i računarstva, Zagreb, Croatia) User Authentication Based on Keystroke Dynamics Analysis  User authentication based solely on password is the most commonly used method, but with several known issues. With the growing number of Internet users, most of whom are not familiar enough with security threats, as well as growing significance of their online activities, stolen passwords are becoming a major security problem. We argue that an additional, non-intrusive level of security can be achieved by analysing keystroke dynamics. A further advantage of the proposed security measure over other biometric measures is that it requires no additional hardware. Tuesday, 5/22/2012 3:00 PM - 8:00 PM, Camelia 2, Hotel Adriatic, Opatija 3:00 PM - 4:00 PM Invited Paper   Ivana Podnar-Žarko (Faculty of Electrical Engineering and Computing, Zagreb, Croatia) Web-tražilice: tehnologije, izazovi i zablude  4:00 PM - 8:00 PM Papers    1. T. Pribanic, I. Culjak, D. Abram, H. Dzapo, M. Cifrek (Faculty of Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia) A brief introduction to OpenCV  The aim of this paper is to introduce OpenCV (Open Source Computer Vision) library. OpenCV is an open source library for image and video analysis, originally introduced more than decade ago. At the beginning OpenCV was an effort made by Intel, aimed to advance research in vision and promote the development of rich, vision-based CPU-intensive applications. Since then, a number of programmers have contributed to the most recent library developments. The latest major change took place in 2009 (OpenCV 2) which includes main changes to the C++ interface. Nowadays the library has >2500 optimized algorithms. It is extensively used around the world, having >2.5M downloads and >40K people in the user group. The purpose of this paper is to quickly make a reader familiar with OpenCV basics without having to go through lengthy reference manuals and books. Of course to master every library element it is necessary to consult many books available on the topic of OpenCV. However, reading such more comprehensive material should be easier after comprehending a basic idea about OpenCV from this paper. Therefore, regardless of whether one is a novice C++ programmer or a professional software developer, unaware of OpenCV, it is certainly advantageous to get briefly acquainted with the major OpenCV building blocks and features. In particular, the main library content should be interesting for the graduate students and researchers in image processing and computer vision areas. 2. L. Mijatović, H. Dean (Sveučilište u Zagrebu, Fakultet elektrotehnike i računarstva, Zagreb, Croatia), M. Rožić (Mikroprojekt d.o.o., Zagreb, Croatia) Implementation of Algorithm for Detection and Correction of Defective Pixels in FPGA  Defect pixels are a common occurence in digital camera sensors, either resulting from the manufacturing process or developing over time. Though low in quantity, they are very noticeable and can destroy the perceived quality of the images. This paper presents a method for detection and correction of defect pixels in images generated by a Bayer mosaic image sensor. We propose an online and adaptive algorithm, which analyzes the images retrieved from a Bayer array sensor on a pixel by pixel basis, considering the values of adjacent pixels to determine if the current pixel is possibly defective, which is either confirmed or refuted by repeating the analysis in subsequent frames. For the confirmed defective pixels, interpolation is performed to restore the image quality. The algorithm is implemented inside an FPGA logic device, suitable for very high frequency operation required to correct defect pixels in images produced by high definition (HD) cameras. 3. I. Grbavac, A. Granić, N. Marangunić (Prirodoslovno-matematički fakultet, Split, Croatia) Emocionalni dizajn: važnost emocija u procesu izrade web stranica  U radu se ističe važnost emocija za proces dizajna web stranica. Pri izradi bilo kojeg predmeta važno je unijeti u dizajn nešto što će kod korisnika izazvati pozitivne emocije. Proizvod koji budi pozitivne emocije pruža užitak korištenja te pozitivna iskustva. Važno je da dizajn svojim izgledom i funkcionalnošću ostvari pozitivna iskustva korištenja kako bi korisnik poželio ponovno koristiti taj ili neki sličan proizvod. Biološki gledano, važno je dizajnom zadovoljiti tri nivoa procesuiranja u mozgu: visceralni, bihevioralni i misaoni. Tek tada dizajn će se smatrati uspješnim. Prvi dio rada opisuje tri nivoa procesuiranja u mozgu, te njihovu vezu s emocijama. Ukazuje se na što je potrebno paziti pri dizajniranju za svaki nivo zasebno. U drugom dijelu rada uspoređuju se dvije web stranice sličnog sadržaja i funkcionalnosti ali različitog dizajna. Pretpostavka je da će web stranica koja je „ljepše“ dizajnirana kod korisnika izazvati više pozitivnih emocija te da će biti procjenjivana lakšom za korištenje, kako kod iskusnih tako i kod manje iskusnih korisnika. Statističkom obradom dobivenih rezultata pokazalo se da je postavljeni cilj istraživanja potvrđen te da se rezultati mogu pripisati utjecaju boljeg ili lošijeg dizajna na procjenu emocionalnog odgovora ispitanika kao i utjecaja na procjenu funkcionalnosti zadanih web stranica. 4. S. Marković, J. Musić, M. Bonković (Fakultet elektrotehnike, strojarstva i brodogradnje Sveučilišta u Splitu, Split, Croatia) Image Based Visual Servoing for Automatic Control of Balancing Beam  While uncalibrated image based visual servoing is traditionally used in robotics, in the paper we propose its application on the problem of automatic control of balancing beam with two thrusters, one at each end. Classical approach to the problem is application of PID regulator or LQR (linear quadratic regulator). Our approach has several important advantages over these classical approaches: knowledge of exact mathematical model is not needed, no issues with observability and controllability, no need for regulator parameter identification, no calibration is needed. In the paper we propose application of Broyden and population algorithms within image based visual servoing framework for control and stabilization of balancing beam. Obtained results are analyzed and compared to results obtained with LQR. Several improvements and possible future research directions are proposed and discussed. 5. E. Dedović, J. Balagija, N. Čavrk, N. Nosović (ETF-Sarajevo, Sarajevo, Bosnia and Herzegovina) Scaling Read-Heavy Web Services  Building scalable applications or improving scalability of existing ones is important and not an easy task to perform. Every application is unique, but they can be categorized in several classes and similar solutions may be applied when scaling applications which belong to the same class. This paper covers five different techniques for scaling read-heavy web services which deliver small amount of data for each request. The techniques analyzed include database scaling, horizontal and vertical architectural scaling, as well as two caching techniques. The service is Ruby on Rails based, but similar approaches may be followed when using other technologies. A comparison between these methods is given, as well as few suggestions related to system bottlenecks and further scaling. 6. S. Krivić, M. Hujdur, A. Mrzic (Elektrotehnički fakultet Sarajevo, Sarajevo, Bosnia and Herzegovina), S. Konjicija (ELEKTROTEHNIČKI FAKULTET, SARAJEVO, Bosnia and Herzegovina) Design and Implementation of Fuzzy Controller on Embedded Computer for Water Level Control  The paper deals with one of frequently encountered tasks in process industry - fluid level control. Proportional Integral Derivative (PID) control is often used for this purpose. Since control parameters of PID controller are fixed and tank system is inherently nonlinear, PID controller should not be used on wider level range. Therefore, this paper analyzes the effectiveness of fluid level control using fuzzy controller. The fuzzy controller is implemented based on mathematical model of tank and using MATLAB. The controller is implemented on Friendly ARM - embedded computer. Arduino board is used as an acquisition board for collecting sensor data from tank system Festo Didactic DD 3100 and as a PWM signal generator for water pump control. Experimental results confirm that the fuzzy control system has good adaptability in comparison with PID and provided satisfying results. 7. A. Fetić, D. Jurić, D. Osmanković (ETF Sarajevo, Sarajevo, Bosnia and Herzegovina) The procedure of a camera calibration using Camera Calibration Toolbox for MATLAB  This paper describes the calibration procedure of CCD digital camera based on the perspective projection method. The idea of this method is to calibrate the camera in order to extract the precise three dimesional information from the images so that they can be used for computer vision and similar applications which take acquired images from the camera as an input. Image capturing proccess reduces the domensions of data from three to two. Camera calibration determines which light source is associated with each pixel on the resulting image. Since the camera used in the process is a complex CCD camera errors resulting from misaligned lenses and deformations in their structures result in more complex distortions in the acquired image. In order to associate points in the camera image space with locations in 3D world space a camera projection matrix consisting of the estimated intrinsic and extrincis parametars is calculated. The calibration process is comprised of capturing calibration images, processing them in order to get the intrinsic and extrinsic camera parameters. The results of calibration are tested using test images captured with the same camera. The calibration process was implemented using MATLAB software package. 8. F. Boltužić, A. Rakipović (Fakultet Elektrotehnike i Računarstva, Zagreb, Croatia) A Hybrid Ant Colony System approach for solving Capacitated Vehicle Routing Problems with Time Windows  The vehicle routing problem with time windows (CVRPTW) represents a common everyday problem. First, the problem is defined mathematically and an example of such a problem is shown. In short, the problem itself is a mixture of the Travelling Salesman Problem and the Bin packing problem. A possible solution to the problem demonstrated, as well as the fitness criteria of the problem. The Ant Colony System Metaheuristic is used to solve the problem. A modification of the original Ant Colony System algorithm is introduced. The heuristics used to improve the basic ACS algorithm are the Savings heuristic, Wait heuristic and the Route size heuristic. Also, a few local search techniques are applied, such as 2*opt local search and Route elimination local search. Finally, the results of the applying the algorithm to a few instances of the problem are discussed. 9. S. Primorac, M. Russo (FESB, Split, Croatia) Android Application for Sending SMS Messages with Speech Recognition Interface  Voice SMS is an application developed in this work that allows a user to record and convert spoken messages into SMS text message. User can send messages to the entered phone number or the number of contact from the phonebook. Speech recognition is done via the Internet, connecting to Google's server. The application is adapted to input messages in English. Used tools are Android SDK and the installation is done on mobile phone with Android operating system. In this article we will give basic features of the speech recognition and used algorithm. Speech recognition for Voice SMS uses a technique based on hidden Markov models (HMM - Hidden Markov Model). It is currently the most successful and most flexible approach to speech recognition. In the future it is planned to elaborate and implement a model that would allow recognition of Croatian words. The application would then have the ability to work without an Internet connection. The process of conversion of speech into text would be carried out exclusively on a mobile device. 10. T. Benussi, D. Jurić (FER, Zagreb, Croatia) A Robust Hand Detection and Tracking Algorithm with Application to Natural User Interface  A natural user interface (NUI) is the term used to describe user interfaces for computers that allow the users to interact with the computer by performing simple movements and gestures. The advantage of such devices is their flexibility due to the lack of spatial constraints from the environment which are present in the use of classic input devices like keyboards and mice. Lately this kind of systems has become very popular and include touch-pads, touchscreens and gaming controllers like the Nintendo Wiimote and the Xbox Kinect. This paper describes a system that implements a natural user interface which allows the user to interact with a computer using hand gestures while being filmed with a web-cam. The proposed system uses a variety of computer vision techniques to accomplish this task and includes a background subtraction algorithm based on histograms of gradients, a skin detection algorithm based on a skin color distribution model and clustering, a hand detection algorithm with morphological features, hand tracking with flocks of features and a hand model based gesture recognition technique. The resulting system shows great robustness to different background and lightning conditions and offers comparable if not superior performance to the benchmark systems of its class. The application of this system has proven to be particularly fruitful in multimedia and desktop environments where it displays a very quick learning period and superior ease of use. 11. M. Brbić, E. Rožić, I. Podnar Žarko (FER, Zagreb, Croatia) Recommendation of YouTube Videos  YouTube is a huge video-sharing service with hundreds of millions of users and hundreds of thousands of videos being uploaded every day. Thus, recommendation of YouTube videos to a single user is a challenging problem which cannot be solved by simply reusing the prevailing recommendation methods. The paper presents a specific recommendation algorithm for YouTube which relies on the data retrieved through the YouTube Data API. A cloud-based application integrates the proposed algorithm and offers a web interface to end users. The paper presents a preliminary analysis of the recommendation quality and lists YouTube Data API limitations which influence the design of recommender systems for YouTube videos. 12. N. Jasika, N. Alispahić, E. Arslanagić, I. Kurtović, E. Lagumdžija, N. Nosović (ETF-Sarajevo, Sarajevo, Bosnia and Herzegovina) Dijkstra's Shortest Path Algorithm Serial and Parallel Execution Performance Analysis  This article introduces the problem of parallelization of Dijkstra's algorithm, a well known algorithm for computing single-source shortest path in a graph. Dijkstra's algorithm can be applied to graphs with a various number of vertices and edges.  Dijkstra's shortest path algorithm is implemented and presented, and the performances of its parallel and serial execution are compared. The algorithm implementation was parallelized using OpenMP (Open Multi-Processing) and OpenCL (Open Computing Language) standards. Its performances were measured on 4 different configurations, based on dual core and i5 processors. The experimental results prove that the parallel execution of the algorithm has good performances in terms of speed-up ratio, when compared to its serial execution. Finally, the results show that, because of Dijkstra's algorithm in itself is sequential, and difficult to parallelize, average speed-up ratio achieved by parallelization is only 8%. This proves to be a huge disadvantage of this algorithm, because its use is widespread, and enhancing its performance would have great effects in its many uses. 13. E. Ajkunić, H. Fatkić, E. Omerović, K. Talić, N. Nosović (ETF-Sarajevo, Sarajevo, Bosnia and Herzegovina) A Comparison of Five Parallel Programming Models for C++  Multi-core processors offer a growing potential of parallelism but pose a challenge of program development for achieving high performance in applications. This paper presents a comparison of the five parallel programming models for implementing parallel programs in C++ on multi-core computer systems. The models under consideration are Intel®’s Thread Building Blocks (TBB), OpenMPI, Intel®'s CilkTM Plus, OpenMP and Pthreads. For demonstration purposes multiple parallel implementations of an algorithm for matrix multiplication suitable for parallelization were created. The main goal of this paper is a comprehensive comparison of chosen models with respect to the following criteria: performance and coding effort required. 14. E. Bajrami, M. Ašić, E. Cogo, D. Trnka, N. Nosović (ETF-Sarajevo, Sarajevo, Bosnia and Herzegovina) Performance Comparison of Simulated Annealing Algorithm Execution on GPU and CPU  This paper presents results of performance comparison of simulated annealing algorithm execution on central processing unit (CPU) and Graphics Processing Unit (GPU) that uses Compute Unified Device Architecture (CUDA). Algorithm used for measurement is a meta-heuristic algorithm based on idea of material crystallization process when temperature is being increased to melting point and then decreased to minimum pre-defined value. Algorithm was implemented for quad-core CPU using C and on CUDA enabled GPU using C for CUDA. Execution time was used as criteria for performance comparison and was measured in two scenarios: fixed number of points with number of threads as a parameter and vice versa. Results of measurements were analyzed and presented. 15. E. Kurtanović, E. Meholjić, N. Hadžić, B. Davidović, E. Ljevaković, N. Nosović (ETF-Sarajevo, Sarajevo, Bosnia and Herzegovina) Parallelization of Array Operations by Using Socket Programming in .NET  This paper describes the procedure and the results of measuring the speed of processing arrays on 1, 2, or 3 hosts, connected in LAN. The measurement scenario was based on having the fixed processing time for each array member in all cases. Also, processing each array member is individual from processing others. For gaining control data, the array was processed on only one host. Afterwards, the array was split in two and three equal pieces for processing on two and three hosts, respectively. The communication between the hosts was established by using socket programming techniques in .NET. The results were impressive because the speed-up in some cases was almost linear. 16. A. Hedji, D. Hrenek, I. Kušt, N. Mikša, P. Prentašić (Fakultet elektrotehnike i računarstva, Zagreb, Croatia) Tiled 360º Panoramic Projection System  In order to obtain 360 degree panoramic displays, tiled video projections are often used. In this paper we present a system for cylindrical 360 degree panoramic projection based on a client-server architecture. The system performs following actions: projector calibration, image transformation and video playback. To perform above actions several problems have to be addressed. Calibration of multiple projectors is based on computing of spatial relations between cylindrical projecting surface and each projector. We do this by photographing the projected chessboard pattern of each projector using a camera with fish-eye lens. The photograph first has to be unwrapped into a rectangle shaped texture. After that, the relation between original chessboard corners and corners on the photograph has to be found. Using these corners the image transformation function has to be interpolated. The transformation of image to be projected is done using this function and also includes lightning correction to achieve smooth transitions between each projector image. If playing video, videos from all projectors have to be synchronized. In the current client-server architecture, a server performs semiautomatic projector calibration, files and configuration distribution to clients and has the ability to off-line create images for each projector. Each client handles storage, showing of already prepared images and playback and real-time transformation of videos. The projector calibration is done using Matlab, whilst the playback and real-time transformation of video is done using custom made software written in C++ and OpenGL 17. D. Petek, D. Komlen, T. Lombarović, A. Petković, M. Ogrizek-Tomaš (FER, Zagreb, Croatia) Multi-Camera Person Tracking System  For the surveillance of the large interior spaces it is often necessary to use multiple cameras. This paper deals with the tracking of an individual using multiple cameras with partially overlapping views. To implement such a system it is necessary to preprocess recordings, detect a person in motion and find the contact point between person and floor. It is especially important to consider the situation in which a person is in the folded multi-camera field of view. Last step is making homographic transformation of a point where person contacts the ground in observed space. System performance was measured on artificially generated recordings of a person moving through halls. It utilizes views with cameras from four different locations. Implemented application runs fast enough for use in systems that operate in real time. 18. D. Hrenek, N. Mikša, P. Prentašić, B. Trubić (Fakultet elektrotehnike i računarstva, Zagreb, Croatia) Artificial Neural Network Based Stock Value Prediction  Financial market turmoil is a new normal today and modern computer science is partialy responsible for this. In this paper we try to show that modern fincancial markets are informationally efficent. In order to show this attribute of financial markets we use a neural network and test it against the S&P 500 stock index. We train our neural network using index information from past in order to predict the future value of the index. We compare the results of neural network based index prediction and a simple buy & hold strategy. Based on this comparison we make a decision about validity of the market efficiency hypothesis. Finally we present some possible improvements to our solution of this problem.

Jelena Ulemek (Croatia), Ines Bombek (Croatia)

Petar Biljanović (Croatia)

Alberto Abello Gamazo (Spain), Slavko Amon (Slovenia), Michael E. Auer (Austria), Mirta Baranović (Croatia), Ladjel Bellatreche (France), Nikola Bogunović (Croatia), Andrea Budin (Croatia), Željko Butković (Croatia), Željka Car (Croatia), Matjaž Colnarič (Slovenia), Alfredo Cuzzocrea (Italy), Marina Čičin-Šain (Croatia), Dragan Čišić (Croatia), Todd Eavis (Canada), Maurizio Ferrari (Italy), Bekim Fetaji (Macedonia), Tihana Galinac Grbac (Croatia), Liljana Gavrilovska (Macedonia), Matteo Golfarelli (Italy), Stjepan Golubić (Croatia), Francesco Gregoretti (Italy), Niko Guid (Slovenia), Yike Guo (United Kingdom), Jaak Henno (Estonia), Ladislav Hluchy (Slovakia), Vlasta Hudek (Croatia), Željko Hutinski (Croatia), Mile Ivanda (Croatia), Hannu Jaakkola (Finland), Robert Jones (Switzerland), Peter Kacsuk (Hungary), Aneta Karaivanova (Bulgaria), Miroslav Karasek (Czech Republic), Bernhard Katzy (Germany), Christian Kittl (Austria), Dragan Knežević (Croatia), Mladen Mauher (Croatia), Branko Mikac (Croatia), Veljko Milutinović (Serbia), Alexandru-Ioan Mincu (Slovenia), Vladimir Mrvoš (Croatia), Jadranko F. Novak (Croatia), Jesus Pardillo (Spain), Nikola Pavešić (Slovenia), Ivan Petrović (Croatia), Radivoje S. Popović (Switzerland), Goran Radić (Croatia), Slobodan Ribarić (Croatia), Karolj Skala (Croatia), Ivanka Sluganović (Croatia), Vanja Smokvina (Croatia), Ninoslav Stojadinović (Serbia), Aleksandar Szabo (Croatia), Laszlo Szirmay-Kalos (Hungary), Dina Šimunić (Croatia), Jadranka Šunde (Australia), Antonio Teixeira (Portugal), Ivana Turčić Prstačić (Croatia), A. Min Tjoa (Austria), Roman Trobec (Slovenia), Walter Ukovich (Italy), Ivan Uroda (Croatia), Mladen Varga (Croatia), Tibor Vámos (Hungary), Boris Vrdoljak (Croatia), Robert Wrembel (Poland), Baldomir Zajc (Slovenia)

Jelena Ulemek
KSET
Unska 3
HR-10000 Zagreb, Croatia

GSM: +385 99 6711 368
E-mail: jelena.ulemek@kset.org

Opatija, often called the Nice of the Adriatic, is one of the most popular tourist resorts in Croatia and a place with the longest tourist tradition on the eastern part of Adriatic coast. Opatija is so attractive that at the end of the 19th and beginning of the 20^th centuries it was visited by the most prominent personalities: Giacomo Puccini, Pietro Mascagni, A. P. Čehov, James Joyce, Isidora Duncan, Beniamino Gigli, Primo Carnera, Emperor Franz Joseph, German Emperor Wilhelm II, Swedish Royal Couple Oscar and Sophia, King George of Greece.

The offer includes 20-odd hotels, a large number of catering establishments, sports and recreational facilities.
For more details please look at www.opatija.hr/ and www.opatija-tourism.hr/.