|
Methodology for identifying the authors behind initials and pseudonyms in the periodical press
Řehák, Daniel
The methodology first deals with the basics behind the issue in question, which are primarily associated with the creation of a journal bibliography. Initially it defines a pseudonym and initials in a general working interpretation and proposes their classification. It then focuses primarily on initials as the main problem in the identification of authors in journal bibliography. Attention is then paid to several levels of difficulty in the deciphering process (manuals, databases, extracts and archive material) and the usability of deciphering for different types of bibliography. The methodology presents the possible practical procedures for the most rapid identification of the connection between the initials and the name of the author using identification techniques and pointing out the option of an inverse procedure for managing larger sets, i.e. a procedure not from the initials to the name, but from the name to the initials. New ways are also proposed to search for information using full-text search engines in scanned newspapers and magazines, particularly with the aim of providing greater detail for data that has already been acquired. Finally a proposal is made regarding the form of the entries for initials and pseudonyms in summary and database records.
Fulltext: content.csg -  PDF
Plný tet: Metodika šifry - PDF
|
|
Analysis of phonation in patients with Parkinson's disease
Kopřiva, Tomáš ; Smékal, Zdeněk (referee) ; Mekyska, Jiří (advisor)
This work deals with analysis of phonation in patients with Parkinson’s disease (PD). Approximately 90% of patients with Parkinson’s disease suffer from speech motor dysfunction called hypokinetic dysarthria. System for Parkinson’s disease analysis from speech signals is proposed and several types of features are examined. Czech Parkinson’s speech database called PARCZ is used for classification. This dataset consists of 84 PD patients and 49 healthy controls. Results are evaluated in two ways. Firstly, features are individually analysed by Spearman correlation, mutual information and Mann-Whitney U test. Classification is based on random forests along with leave-one-out validation. Secondly, SFFS algorithm is employed for feature selection in order to get the best classification result. Proposed system is tested for each gender individually and both genders together as well. Best result for both genders together is expressed by accuracy 89,47 %, sensitivity 91,67% and specificity 85,71 %. Results of this work showed that the most important vowel realizations for phonation analysis are sustained vowels pronounced with maximum or minimum intensity (not whispering).
|
|
Reasons for NoSQL deployment, and especially MongoDB
Čermák, Miroslav ; Palovská, Helena (advisor) ; Tomášková, Barbora (referee)
This bachelor thesis examines reasons of NoSQL databases deployment and especially MongoDB database. It introduces these databases, describes its classification, specific properties, including its advantages. It continues with introduction of MongoDB and its case studies. Advantages of this database are shown on the custom model of e-commerce system. These reasons are then summarized in the individual chapter.
|
|
High temporal satellite data assimilation for vegetation spectral characteristic assignment
Malíková, Lucie ; Langhammer, Jakub (referee) ; Štych, Přemysl (advisor)
The application of high temporal satellite image data for designation of the spectral characteristic of vegetation Abstract The objektive of this paper is to evaluate possibilities of high temporal satellite data assimilation for continuous monitoring of the spectral characteristic of vegetation. There is also given the suggestion of metodology for processing MERIS data and for continuous monitoring of spectral characteristic of landscape objects. Finally, vegetation cover database for the Czech Republic in the year 2009 is created from sectorial analysis. In the paper there is used the LSU classification and thresholding of vegetation indicies histograms. The universal decision algorithm for classification of vegetation landscape component are described and particular thresholding values for the year 2009 given. The finally product of this paper is Czech vegetation cover database for the year 2009 with overall accuracy of 63,35 %. Accuracy for forest is even over 80 %. Keywords: remote sensing, MERIS, vegetation, spectral reflectance, LSU, BEAM
|
|
Software for manual delineation of ECG signals
Jež, Radek ; Kozumplík, Jiří (referee) ; Vítek, Martin (advisor)
This thesis deals with evaluation EKG in terms of classification rhythm and analysis HRV. In theoretic part of work are described basics of heart physiology and its usual pathology, basics of electrocardiography, evaluation EKG and standard methods of HRV evaluation. In practical part are described algorithms used in created application. Mainly describes technique of rhythm evaluation, ectopic rhythms and delineation error elimination, data preparing for HRV evaluation, drift removal from DES and HRV evaluation methods. Created program was tested on CSE and MIT- BIH database records. For lack of suitable data and absence of tested data, it wasn’t possible to test all the classification rules of used algorithms. Tested part of program appears reliable and functional.
|
|
Texture-Based Object Recognition
Hutárek, Jiří ; Švub, Miroslav (referee) ; Španěl, Michal (advisor)
Main subjects of this thesis are texture classification and texture-based object recognition. Various texture features are being explored, including several variants of local binary patterns (LBP). A novel modification of LBP (weighted spatial LBP) is proposed, with intention to improve on the spatial coverage of the traditional LBP. Rarely used color texture features are being discussed as well. Artificial neural networks and support vector machines are used to classify all the aforementioned features. Using these methods, framework for the texture classification and image segmentation is implemented. Comprehensive texture database is employed to test its performance under different conditions. In the end, the system is applied to solve a real-world problem - the segmentation of aerial photos.
|
|
Copyright protection of databases
Hájková, Martina ; Dobřichovský, Tomáš (advisor) ; Císařová, Zuzana (referee)
The topic of this diploma thesis is copyright protection of databases. The aim of the thesis is to describe and introduce databases and their legal protection due to their role as essential tools for the location of required information. The thesis contains analysis of legislation in European Union and the Czech Republic and includes brief analysis of legislation in Finland. The thesis consists of five chapters. The first chapter deals with the term database, classification of databases and illustrates basic foundations for database protection. The second chapter is divided into two subchapters and characterises the legal protection of databases in European Union. The first subchapter is concerned with the legislation process and adoption of the directive on the legal protection of databases starting with the Green Paper in 1988 and ending with adoption of the directive in 1996. The second subchapter analyses this directive in more detail, describes the main terms and institutes and explores the two-tier system of legal protection of databases introduced by this directive. The third chapter presents legislation in Finland with regard to the exchange studies of the author of the thesis and outlines the catalogue rule as a traditional element of Scandinavian law. The fourth chapter is divided into...
|
|
Hard and soft exudates detection in retinal images
Válková, Hana ; Lamoš, Martin (referee) ; Kolář, Radim (advisor)
The thesis deals with automatic detection of soft and hard exudates in retinal images of the human eye. In its introduction the thesis describes the issue of diabetes in relation to the damage to the retina of the eye. What is described in the first place is diabetic retinopathy, its symptoms and progression of the disease. Another section is devoted to describing DIARETDB1, the freely accessible database which besides other things contains a set of images showing various degrees of disease, evaluation of images from the experts and the evaluation protocol. The next section discusses several methods for automatic detection of hard and soft exudates. The practical part of the bachelor’s thesis is aimed at image pre-processing with respect to the normalization of retinal images, the selected method for adaptive transformation of contrast was implemented. This part also containts description of chosen methology of thresholding, feature extraction based on lesions intensity and its surroundings, use of Ho Kashyap classifier is described, classification of lesions in images is followed. In conclusion realized methods is evaluated.
|
|
Modification of company processes supported by IT
Slámová, Pavla ; Říha, Tomáš (referee) ; Kříž, Jiří (advisor)
The aim of this thesis is to suggest and implement a multi-purpose database of orders in advertising agency, which will uniform and make transparent the process of giving orders to individual suppliers and will enable its overall administration, survey and classification of all assigned orders. The implementation of the database system leads to closure of the gap in agency processes, where the incomplete and not always available summary of all the outbound orders complicates the running and final invoicing of the individual orders to clients.
|
|
Classifiaction algorithms for face-based identification systems
Hegr, Vojtěch ; Křupka, Aleš (referee) ; Malach, Tobiáš (advisor)
The thesis deals with the research of classification algorithms for face-based identification. The aim is to implement algorithms into an existing system for face recognition and the evaluation of the impect of individual classifiers. According to the survey of face recognition methods the following classifiers were chosen for implementation: K - Nearest Neighbours (K-NN), Support Vector Machines (SVM) and the Neural Networks. These classification algorithms were implemented in C++ (Microsoft Visual Studio 2010) using the open source library OpenCV. Furthermore, the IFaVID database and the methodology used to test the implemented algorithms were introduced.
|
guest ::
