|
|
Automatic Machine Learning Methods for Multimedia Data Analysis
Mašek, Jan ; Chromý, Erik (referee) ; Vozňák, Miroslav (referee) ; Burget, Radim (advisor)
The quality and efficient processing of increasing amount of multimedia data is nowadays becoming increasingly needed to obtain some knowledge of this data. The thesis deals with a research, implementation, optimization and the experimental verification of automatic machine learning methods for multimedia data analysis. Created approach achieves higher accuracy in comparison with common methods, when applied on selected examples. Selected results were published in journals with impact factor [1, 2]. For these reasons special parallel computing methods were created in this work. These methods use massively parallel hardware to save electric energy and computing time and for achieving better result while solving problems. Computations which usually take days can be computed in minutes using new optimized methods. The functionality of created methods was verified on selected problems: artery detection from ultrasound images with further classifying of artery disease, the buildings detection from aerial images for obtaining geographical coordinates, the detection of materials contained in meteorite from CT images, the processing of huge databases of structured data, the classification of metallurgical materials with using laser induced breakdown spectroscopy and the automatic classification of emotions from texts.
|
|
|
Signal processing using parallel mathematical operations
Polášek, Jaromír ; Ležák, Petr (referee) ; Mžourek, Zdeněk (advisor)
This Bachelor thesis deals with the acceleration of function calculations, using parallel computing mediated by NVDIA graphics cards via CUDA technology. The theoretical part describes the general principles of parallel computing and the basic characteristics and parameters of graphics cards NVDIA. The theoretical part also deals with basic principles of CUDA technology. End of the theoretical part focuses on FFTW and cuFFT libraries. The practical part deals with the comparison of the performance between GPU and CPU functions filter2D and Canny and practical possibilities of accelerating fast convolution calculation. The practical part also describes sample code that was used to compare the performance between GPU and CPU. The results of this program are then plotted and evaluated.
|
|
|
Use of optical cable properties for geolocation in the Internet
Chmelař, Jakub ; Verner, Lukáš (referee) ; Komosný, Dan (advisor)
This thesis deals with the IP geolocation which estimates the geographic location of hosts in the Internet network. The knowledge of geographic location can be used both for targeting specific services and information, such as advertisement or weather forecast, and for better utilization of transport network because it is possible to serve content from sources which are located closer to client. The main topic of this thesis is the Speed of Internet geolocation method (SOI), which is an active method based on measurement of latency in network. The theoretical part of the thesis explains how to perform these measurements and briefly describes a few more active geolocation methods. The practical part includes full description of SOI implementation created using Python programming language. This implementation was subsequently used to accomplish a sequence of measurements which allowed accuracy comparison with other geolocation methods. The final part of this thesis evaluates results of these measurements and offers a few possible changes which could improve accuracy of described implementation.
|
|
|
Image Processing in FPGA
Maršík, Lukáš ; Španěl, Michal (referee) ; Zemčík, Pavel (advisor)
This bachelor's thesis presents a hardware realization of graphic algorithm for rendering objects described with 3D point clouds - a spatial objects representation. An FPGA (Field-Programmable Gate Array) chip coupled with a DSP (Digital Signal Processor) creates basement for implementation of function units. Is possible to decrease overall computation time by using more than one of that pair. That mean so simple distribution of load is used. The input graphical data is 3D point clouds - sets of points which are transformed into oriented circles just for purpose of rendering. Result of projection of that elements are ellipses. Such graphical representation seems to be more suitable for many purposes than the most commonly used triangle meshes. The implementation equivalent to concept is described too.
|
|
|
GPU Image Processing Library
Čermák, Michal ; Španěl, Michal (referee) ; Smrž, Pavel (advisor)
This work is concerned with architecture of recent Nvidia graphics cards and application programming interface CUDA. That is used to create accelerated image processing library. It place emphasis on testing performance gain compassion with high optimized and used OpenCv library.
|
|
|
Workflow Model and Its Graphic Interface
Jadrný, Miroslav ; Mates, Vojtěch (referee) ; Hruška, Tomáš (advisor)
Business process management is important topic in business information systems. Workflow systems are taking the top places in company information system architecture due to aspiration to make business process more and more optimized. This project is about parallel processing and implementation of business processes parallel processing in complex information systems. Content of this project is o function and object library for modeling business process in Vema, a. s. Workflow system. Important part of this project is parallel processing solution and its implementation.
|
|
|
Distributed Ray Tracing in Reasonable Time
Slovák, Radek ; Polok, Lukáš (referee) ; Herout, Adam (advisor)
This thesis deals with the method of distributed ray tracing focusing on optimalization of this method. The method uses simulation of some attributes of light by distributing rays of lights and it produces high quality and partly realistic images. The price for realitic effects is the high computational complexity of the method. The thesis analysis the theory connected with these aspects. A large part describes optimalizations of this method, i.e. searching for the nearest triangle intersection using kd-trees, quasi random sampling with faster convergence, the use of SSE instruction set and fast ray - triangle intersection. These optimalizations brought a noticable speed - up. The thesis includes description of implementation of these techniques. The implementation itself emphasises the practical usability including generating some advanced animations and universal description of objects.
|
|
|
Database Performance Tuning
Paulíček, Martin ; Ruttkay, Ladislav (referee) ; Chmelař, Petr (advisor)
The objective of this thesis was to study problems of an insufficient database processing performance and possibilities how to improve the performance with database configuration file optimizations, more powerful hardware and parallel processing. The master thesis contains a description of relational databases, storage media and different forms of parallelism with its use in database systems. There is a description of the developed software for testing database performance. The program was used for testing several database configuration files, various hardware, different database systems (PostgreSQL, Oracle) and advantages of parallel method "partitioning". Test reports and evaluation results are described at the end of the thesis.
|
|
|
Acceleration of Discrete Optimization Heuristics Using GPU
Pecháček, Václav ; Jaroš, Jiří (referee) ; Pospíchal, Petr (advisor)
Thesis deals with discrete optimization problems. It focusses on faster ways to find good solutions by means of heuristics and parallel processing. Based on ant colony optimization (ACO) algorithm coupled with k-optimization local search approach, it aims at massively parallel computing on graphics processors provided by Nvidia CUDA platform. Well-known travelling salesman problem (TSP) is used as a case study. Solution is based on dividing task into subproblems using tour-based partitioning, parallel processing of distinct parts and their consecutive recombination. Provided parallel code can perform computation more than seventeen times faster than the sequential version.
|
|
|
Optimization of ETL processes
Huňa, Tomáš ; Pour, Jan (advisor) ; Novotný, Ota (referee)
This bachelor's thesis deals with optimization of ETL processes; it is extraction, transformation and load of data warehouse. It is about performance optimization of ETL. The aim of the thesis is to explore ways the particular ETL phases can be optimized, to describe recommendations, methods, suggestions and findings, which can positively affect performance, and to summarize everything into one piece in Czech. The thesis is divided into three chapters. The first chapter deals with basic description of ETL, ways how to estimate duration of processing, and ways how to measure performance of each phase of ETL. Next part is aimed at optimization. Suggestions, methods and recommendations are introduced both for whole ETL and for every single process; extraction, transformation and load. The last chapter is about parallel processing, which can be applied to all ETL processes.
|
guest ::
