|
|
Person Identification and Verification Using EEG
Žitný, Roland ; Orság, Filip (referee) ; Tinka, Jan (advisor)
The aim of this work was to create a brain-computer interface that reliably identifies and verifies a person using his electroencephalographic signals. Creating a user profile and verifying it is based on processing reactions to his own face, and the face of strangers or acquaintances. Algorithms such as bandpass and noise removal using wavelet transformation are user to filter signals. The classification of reactions is performed using a convolutional neural network or linear discriminant analysis. The average accuracy of the linear discriminant analysis is 66.2 % and of the convolutional neural network is 58.7 %. The maximum achieved accuracy was with linear discriminant analysis and at 93.7 %.
|
|
|
Detection of Landscape in Images
Dufka, Zbyněk ; Křupka, Aleš (referee) ; Číka, Petr (advisor)
The target of this thesis is theoretical summary of methods, which are used for detection of objects in database of images. Next step is to develope own algorhytm for detection of landscape. Theoretical part describes methods Viola-Jones and HOG, which are used for detection of human faces and anomalies on radiographs. This analysis is concerned with segmentation method and algorhytm for data mining from images, which contain the landscape scenery. Appropriate method and its analysis could contribute to create and to implement into developement enviroment RapidMiner in the second part of thesis.
|
|
|
Automated Detection of graphic elements in EEG signal
Jančová, Ivana ; Ronzhina, Marina (referee) ; Bubník, Karel (advisor)
This thesis deals with the analysis of EEG, namely detection of graphic elements. The aim of the thesis is to describe methods suitable for detection of graphic elements and implementation of the two methods in MATLAB. The first part of the thesis describes the basics rhythms, artifacts and epileptiform transients occurring in the EEG. The second part deals with the methods of detection and their mutual comparison. The discussed methods are spectral analysis, correlation analysis, wavelet transform, cluster analysis and matched filter. Other parts of the thesis describe the database PhysioBank and implementation of the correlation analysis and matched filter. In the last part is the comparison of the success of selected methods and comparison is done by calculating sensitivity and positive predictive value.
|
|
|
ECG quality estimation
Pospíšil, Jan ; Smíšek, Radovan (referee) ; Smital, Lukáš (advisor)
This bachelor thesis deals with the question of estimation of the quality of the ECG signals, which is a key parameter for determining the diagnosis. The theoretical part deals with the basic knowledge concerning cardiac physiology, electrocardiography and finally the types of interferences that can occur during the measurement. The following practical part will deal with the published methods and the proposal of methods for estimating signal quality and their testing on artificial and real data.
|
|
|
Comparison of Modern Lossy Image Compression Formats
Kalabza, Marek ; Klíma, Ondřej (referee) ; Bařina, David (advisor)
This thesis compares modern compression methods of digital images on given datasets in terms of quality. First part describes the principle of lossy compression and then compression of JPEG, Lepton, JPEG 2000, JPEG XR, WebP, BPG, HEIF, H.265 intra frame, H.264 intra frame and AV1. Second part focuses on testing, comparison of results and evaluation. It also describes quality metrics (PSNR, PSNR-YUV, SSIM), deals with possible artefacts created during compression and points out the behavior of formats after data corruption. Based on testing and measurement, it was found that the best compression format from the tested lossy image formats is the AV1 format, specifically on the high resolution dataset, the scanned document dataset, and the map dataset.
|
|
|
Effect of artefacts on determination of aerobic threshold
Jaroš, Oliver ; Korvas, Pavel (referee) ; Janoušek, Oto (advisor)
The bachelor's thesis deals with the identification of artifacts occurring during running and swimming and the influence of these artifacts in determining the aerobic threshold from RR intervals using the analysis of detrended fluctuation and its 1 index during a stress test. First, the theoretical part is discussed, where all information related to the definition of training zones, HRV, detrended fluctuation analysis and the method of determining the aerobic threshold from DFA are described. The second part of the work is practical and deals with specific measurements, their processing and subsequent statistical analysis of the results. It was found that for artifacts occurring during running, the mean absolute error in determining the aerobic threshold was only 1.58 beats/min. and after applying the correction method only 0.33 beats/min. When analyzed with artifacts occurring during swimming, this error was up to 9.89 beats/min, and after applying the correction method it was only 0.67 beats/min.
|
|
|
Product Data Visualization in PNC Build System
Korytár, Patrik ; Janoušek, Vladimír (referee) ; Kočí, Radek (advisor)
Cieľom tejto bakalárskej práce je zlepšiť vizualizáciu dát súvisiacich s produktami build systému PNC na jeho novom webovom používateľskom rozhraní. Hlavné vizualizačné prvky zahŕňajú tabuľky, grafy, sieťové grafy a palubné panely. Táto práca analyzuje vstupy od používateľov build systému PNC, aby sa zohľadnili skutočné potreby používateľov. Na základe analýzy bol vytvorený dizajn pre úplne nové komponenty používateľského rozhrania a niektoré z už existujúcich boli prepracované. Na ilustráciu nového dizajnu sa vytvorili nákresy. Navrhnuté boli nové koncové body REST API pre dáta vyžadované novými komponentami. Nové funkcie boli následne implementované. Implementačný jazyk je TypeScript. Hlavné knižnice zahŕňajú React, Chart.js, Sigma.js a Graphology. Ako výsledok má teraz nové používateľské rozhranie systému PNC nové funkcie, ktoré pomáhajú s používaním stránok súvisiacich s produktami. Implementácia je tiež ľahko rozšíriteľná pre viac vizualizácií tohto druhu.
|
|
|
Clothing as a cultural artifact and its current use in art lectures at primary education
Melounek, Robin ; Kovandová, Anna (advisor) ; Jakubcová Hajdušková, Lucie (referee)
Melounek, R.: Clothing as a cultural artifact and its current use in Art lectures at Primary Education [Thesis] Prague 2013 - Charles University, Faculty of Education, Department of Art Education, 113 p. Theoretic part of thesis deals with finding the mutual relations between the history of the 20th century clothing with wiew on combination of clothes, rituals and art. Didactic part offers a preview of current social rituals, through children's artistic perception, of which we may experience in life and personally reflect. In the practical section, through creation of a unique garment, thesis examine the possibilitie of understanding of clothing as a work of art and it's dual usage.
|
|
|
Person Identification and Verification Using EEG
Žitný, Roland ; Orság, Filip (referee) ; Tinka, Jan (advisor)
The aim of this work was to create a brain-computer interface that reliably identifies and verifies a person using his electroencephalographic signals. Creating a user profile and verifying it is based on processing reactions to his own face, and the face of strangers or acquaintances. Algorithms such as bandpass and noise removal using wavelet transformation are user to filter signals. The classification of reactions is performed using a convolutional neural network or linear discriminant analysis. The average accuracy of the linear discriminant analysis is 66.2 % and of the convolutional neural network is 58.7 %. The maximum achieved accuracy was with linear discriminant analysis and at 93.7 %.
|
|
|
Noise and artifact suppression in fMRI data based on multi-echo data and independent component analysis
Pospíšil, Jan ; Gajdoš, Martin (referee) ; Mikl, Michal (advisor)
The main task of this work is to design an algorithm for suppressing unwanted noise and artifacts in fMRI data using the analysis of independent components and multi-echo data. The theoretical part deals with the basic principles of magnetic resonance, including construction and image data processing. The practical part presents a pilot design of a method inspired by a professional publication in the Matlab software environment, where this design is subsequently tested on real fMRI data provided by the Laboratory of Multimodal and Functional Imaging, CEITEC MU.
|
guest ::
