Monthly Archives: May 2016

FIPA Standardization for Agent Based Technology

Georgi Tsochev, Roumen Trifonov, Georgi Popov

Abstract: When developing new agent applications there are curtain standards that must be observed. This paper reviews the some of the most commonly used standards. The main aspect is the brief review of the Foundation for Intelligent Physical Agents (FIPA) standard.

Keywords: agent systems, security intelligence, standardization, fipa, software agent

Management and Manipulation of Knowledge in Automated Learning Systems

Ventzislav Nikolov, Stephan Filipov, Ivan Gospodinov

Abstract: The subject of this paper is the use of production systems for presentation and manipulation of knowledge with constant and variable structure in automated educational systems. A TOMt approach for construction of production rules is proposed. The non-excess and non-contradiction of the mles are studied. An algorithm for construction of correctly built system of production of knowledge of variable structure of the type <condition -> recommendation> is created. Relation tables are used as structures reflecting the facts of the subject learning domain, whereas a dynamic tree structure /graph/ is built for describing the semantics of the subject domain. The mechanism that governs the learning process is based on the production systems.

Keywords: relational databases, dynamic trees, production rules, production systems, software engineering, knowledge engineering

Download article 

Experimental Studies on the Performance of Algorithms for CFA Interpolation

Iva Nikolova

Abstract: This paper deals with the challenging task of reconstructing a full three-color representation of color images captured by a single image sensor – CCD or CMOS, which samples the image according to a color filter array (CFA). The output images feature a certain kind of pattern in which every pixel has only single-color information. The process of modifying the raw image sensor data to obtain full RGB color information at each pixel is called demosaicing or CFA interpolation. Seven popular algorithms CFA interpolation are regarded. The results of timing performance analysis of their sequential implementations are presented and discussed. The performance speedup analysis and scalability tests with multithreading and CUDA-based GPU implementations are analyzed.
Keywords: Bayer pattern, CFA interpolation, GPU, Image sensors, Multithreading, Parallel computations, Performance and speedup analysis.

Download article

Virtualization Technologies for Education on Computer Networking and Research

Mitko Shopov

Abstract: The use of virtualized platforms is a cost effective and affordable solution. Virtualized environment offers high flexibility in laboratory experiment design as topology changes can be made by software changes. The high scalability and flexibility of virtual networking enables design of large scale, complex, and advanced networking experiments and laboratory exercises that would not be built otherwise using physical equipment. The paper presents the use of virtual networking technologies along with GNS3 tool in the education and research on computer networking in DSNET virtual laboratory, part of Technical University of Sofia, branch Plovdiv.
Key words: Computer Networking, GNS3, Qemu, Virtualization, VMWare

Download article

Sensor Data Processing — Key Factors for Choosing File Formats in HADOOP

Stefan Stoyanov

Abstract: Nowadays, in some areas, data growth has reached the point where a single relational database is not enough. This big data phenomenon has first time appeared in areas like meteorology, sensor data analytics, Internet search, biological research, genomics, finance, and many more. The effective use of distributed resources in clusters or data warehouses depends on the selection of appropriate data format and best algorithms. The subject of this research is a methodology and approaches for optimal use of the resources of a distributed system for processing Big Data from sensor measurements.

Key words: Big Data, Hadoop, HDFS, Sensor Fusion

Download article 

Evaluation of the Routing Algorithms with Reduced Energy Consumption

Ile Dimitrievski, Valentin Molloy

Abstract: This paper presents an evaluation of a special routing algorithm with energy reduced consumption in comparison to one without any energy consumption optimization. We apply Noxim discrete event simulator specialized for simulation of Network on Chip (No C) architectures. Evaluation is made over a 7×7 2D mesh structure which is the most used model nowadays. Comparison between the popular state-of-the-art XY routing algorithm and DyAD one is given, as well. DyAD routing algorithm for NoC combines the advantages of both deterministic and adaptive routing algorithms. Calculations on the above algorithms are performed by means of the Noxim simulator with examples on 2D mesh topology.

Key words: Adaptive, deterministic, DyAD, Network on Chip, power consumption, random selection, routing algorithms, 2D mesh topology.

Download Article

A Model of Hybrid Ontology for Medical Information

Mariya Evtimova, Ivan Momtchev

Abstract: It is proposed a model of hybrid ontology that is case based, but is also suitable for big data as it is implemented rules. This ontology is divided into three parts- crisp part, fuzzy and probability part and big data part. Ontology use theory of the fuzzy sets with addition of probability logic for the realization of the fuzzy and probability part of the ontology. This hybrid ontology is suitable for vague and uncertain reasoning as it improve the quality of the returned results. Quality of the returned results is very important in the medical domain.

Key words: Big data, Case based reasoning, Fuzzy logic, Ontology, Probability logic

Download Article

A Hybrid Technique for Kidneys Segmentation from Axial CT Images

Elena Kalcheva-Yovkova, Technical University, Sofia

Abstract: This paper represents some results of a hybrid technique for kidney segmentation from axial computed tomographic images. The proposed technique combines edge based, region based and morphological watershed based algorithms with prior anatomical knowledge of the human abdomen. It includes kidney region extraction, coarse watershed based segmentation and adaptive refinement. The approach is tested on several CT datasets in DlCOM standard. The results of the segmentation of the left and right kidneys are quantitatively assessed using supervised evaluation technique including similarity, consistency and distance measures. The average values of specificity, sensitivity and accuracy are relatively high (above 0,96) and mean value of Dice coefficient is about 0, 95. The global consistency error values alter between 0,02 and 0,05 and the average Hausdorff distance is around 0,06.

Keywords: Computerized tomography images, contrast enhancement, morphological gradient, watershed segmentation, supervised evaluation.

Download Article