Boris Rachev

Multidimensional models: constructing data CUBE

The goal of this article is to depict algorithms and new approach of creating data cube efficiently and to set tasks for future work. In short, the paper has described it as a data abstraction that allows one to view aggregated data from a number of perspectives: 1) Conceptually, the cube consists of a base cuboid, surrounded by a collection of sub-cubes that represent the aggregation of the base cuboid along one or more dimensions; 2) The basic methods for computing a group-by: sort-based and hash-based. 3) Dynamic Data Cube, which provide efficient performance and allows for the data cube expansion in any direction. 4) OO Conceptual Model Data Cube - class definition of data cube.

An approach for design of the object-oriented databases with restricted natural language access

The paper proposes a three-dimensional matrix model for representation of entities, attributes, relationships and attribute values of databases in order to access them through menus with predefined fields when posing queries.It is discussed that the databases, described by the proposed model can be synonymously defined by the structures, used in the modern object-oriented programming languages, such as classes, objects and dynamical arrays.

Image databases: an approach for image segmentation & color reduction analysis & synthesis

In this paper, we describe a system for image databases organization and retrieval based on the pictorial content representation. We suggest extending the use of image histograms to characterize the global and local color properties of an image and examine the effect of the degree of image segmentation and a number of known color reduction algorithms on the time taken for image database creation and retrieval and on the size of the metadata, extracted from the images. We drew the following conclusions from the analysis of the results of our experiments: the color reduction algorithms have a relatively small effect while the size of the database and the times taken for creation and retrieval increase significantly if the image segmentation is above 16×16. We have also concluded that this type of structure should be based on a hierarchical method of segmentation.

CompSysTech: ten years after the beginning...

The Conference on Computer Systems and Technologies (CompSysTech) was initiated by the Bulgarian ACM Chapter (acmbul) in collaboration with the Academic Society of Computer Systems and Technologies and started in 2000 as a mainly Bulgarian National Event in Computer Systems and Technologies. Later, into consortium was included the Bulgarian Union of Automatics and Informatics (UAI). Since then, CompSysTech has been morphed into an internationally visible scientific forum in the area of Computing. During the last ten years the key aim of CompSysTech series of Conferences is to present and diffuse research results into the research communities of Balkan Region and Europe Countries and to increase the international scientific collaboration. This paper presents main statistic and tendencies of the CompSysTech Events Progress. The raw data have been collected from the CompSysTech proceedings and CompSysTech Paper and Review System Database.

An approach for texture SDI visualization

In this paper we propose an algorithm for generating synthetic texture with desired appearance and size. Also, we propose a proper visualization process for using the received texture to obtain a realistic appearance of 3D surfaces which can by applied in geographic maps, for mapping of a vegetation and other special features onto terrains, to correct failings of a photo, video and satellite images.