|
| |
|
|
Visualization and testing of team profiles
Höschl, Cyril ; Mráz, František (advisor) ; Holan, Tomáš (referee)
The aim of this thesis is to demonstrate methods for visualization and testing team profiles. An individual's profile is a vector of scores representing personality of a team member. Team profile is a collection of individual profiles. The first step for visualization of team profile is to find similarities among individual profiles. We present new approach for measuring similarity between two profiles as a significance of difference expressed by probability. We use this measure both for testing mean difference between two subgroups of profiles and also for building similarity matrix. We represent team profiles as a 3D ladscape map called Sociomap. To create a sociomap we demonstrate own algorithm for projecting similarity matrix into a 2D plane. We compare this algorithm with well known Multidimensional scaling and Principal Component Analysis methods solving the same task. Comparing with MDS and PCA, our algorithm provides more accurate result with comparable demand of time. We also employ some well known Computer Graphics methods to improve visual 3D representation of the sociomap and we also show some new improvements to render sociomap more realistic.
|
|
|
Development of photic entrainment of the circadian clock of the rat during early postnatal ontogenesis
Matějů, Kristýna ; Sumová, Alena (advisor) ; Höschl, Cyril (referee) ; Langmeier, Miloš (referee) ; Nevšímalová, Soňa (referee)
In most organisms, behavioral and physiological events oscillate with period ± 24 h, i.e. exhibit circadian rhythms. In mammals, circadian rhythms are generated by circadian clock within the suprachiasmatic nuclei of the hypothalamus (SCN). Light entrains circadian rhythms to the 24 h period of solar day. Information about light is conveyed from the retina to the SCN and induces expression of clock genes Period1 (Per1) and Period2 (Per2) that represent photosensitive parts of molecular circadian clockwork within SCN. Light sensitivity of Per1 and Per2 within SCN is temporally restricted to the subjective night phase. In addition, daily profiles of clock gene expression within SCN are modulated by daylenght, i. e. the photoperiod. The aim of our study was to elucidate how the mechanism of photic entrainment of the rat circadian clock develops during prenatal and early postnatal ontogenesis. Our results demonstrate that maternal circadian system provides information about external light to the fetal and early postnatal circadian clock. Circadian clock within the SCN of rat pups is light sensitive already at the first postnatal day. Mechanism gating the light sensitivity is present at postnatal day 3 and gradually maturates until postnatal day 10. The data suggest that the developing retina is responsive to...
|
|
|
Sociomap Visualization
Höschl, Cyril ; Bálek, Martin (advisor) ; Černý, Jakub (referee)
In this work we study algorithms for visualization of sociomaps. We focus on surface-fitting method which converts bitmap picture into a sequence of closed spline curves. For curve-fitting, we describe tracing algorithm that for a given level of height in a smooth surface finds a contour as a closed Bézier curve. Further we describe algorithm that draws hierarchy of clusters into a 2D plane. We first create a skeleton of the cluster analysis which we then project into a plane using curve-fitting method. We describe methods of interpolation of surface based on inverse distances weights and also we present a linear interpolation based on Delanauy triangulation. We apply these methods on sociomappig analysis. Sociomapping is a data-mining method for visualization of relations among objects as well as analyzing statistical data. The result of sociomapping analysis is an easy-to-read picture - a sociomap. We explain cluster analysis for model of a sociomap and also we describe statistical test based on fuzzy approach.
|
|
|
Advanced Moment-Based Methods for Image Analysis
Höschl, Cyril ; Flusser, Jan (advisor) ; Papakostas, George (referee) ; Jiřík, Radovan (referee)
The Thesis consists of an introduction and four papers that contribute to the research of image moments and moment invariants. The first two papers focus on rectangular decomposition algorithms that rapidly speed up the moment calculations. The other two papers present a design of new moment invariants. We present a comparative study of cutting edge methods for the decomposition of 2D binary images, including original implementations of all the methods. For 3D binary images, finding the optimal decomposition is an NP-complete problem, hence a polynomial-time heuristic needs to be developed. We propose a sub-optimal algorithm that outperforms other state of the art approximations. Additionally, we propose a new form of blur invariants that are derived by means of projection operators in a Fourier domain, which improves mainly the discrimination power of the features. Furthermore, we propose new moment-based features that are tolerant to additive Gaussian image noise and we show by extensive image retrieval experiments that the proposed features are robust and outperform other commonly used methods.
|
|
|
Bipolar Disorder and Diabetes Mellitus
Růžičková, Martina ; Höschl, Cyril (advisor) ; Haluzík, Martin (referee) ; Papežová, Hana (referee)
Despite lots of research, the genetic and pathophysiological basis of bipolar disorder (BD) remains unclear. One of the possible reasons is that BD spectrum comprises a heterogeneous group of different subtypes of the illness sharing certain pathophysiological or genetic mechanisms. Patients with co-morbid diabetes mellitus (DM) may represent such a subgroup of BD with a distinct pathophysiology and possibly different clinical characteristics of BD. This thesis is aimed at investigating the possible link between BD and DM. Bipolar patients have about 3 times higher risk of DM (type 2 in particular). The first part of my thesis outlines the possible links between DM and BD, including medication, alterations in the hypothalamo-pituitary axis and signal transduction, and genetics. In the second part, I examined the possible differences in clinical characteristics between patients with (N=26) and without (N=196) co-morbid DM. Next, I examined the variables showing differences between groups, using logistic regression. Patients were recruited via The Maritime Bipolar Registry. The prevalence of DM was 11.7%. Diabetic patients were significantly older than non-diabetic patients (P < 0.001), had higher rates of rapid cycling (P = 0.02), more chronic course of BD (P = 0.006), more disability (P < 0.001), lower...
|
|
| |
|
|
New Neurophysiological Indicators Measuring, Monitoring and Evaluating in Selected Psychiatric Disorders
Šóš, Peter ; Höschl, Cyril (advisor) ; Libiger, Jan (referee) ; Hájek, Milan (referee)
There is increasing amount of new knowledge in neuropsychiatric research, partially as a result of application of new neurophysiological methods of quantitative electroencephalography. This is because these techniques have good temporal resolution, are relatively non-invasive and economical. Different experimental paradigms and electroencephalographical analytical techniques are reviewed, with their potential applications in the assessment of psychiatric disorders. In the theoretical part the principles of some methods of quantitative electroencephalography, especially are described: EEG coherences, QEEG cordance and LORETA. Next chapter contains basic information about panic disorder and major depressive disorder followed by findings in functional imaging studies which are associated with the selected psychiatric disorders. Separate section is devoted to the prediction of antidepressant response. The empirical part of the thesis consists of a set of the five original papers about application of QEEG methods in a population of psychiatric patients (Bareš et al., 2010; Kopeček et al., 2008; Kopřivová et al., 2009; Šóš et al., 2008; Šóš et al., 2013). In Study 1, we followed the localization of EEG sources by means of brain electromagnetic tomography with low resolution (LORETA) in patients with panic...
|
|
|
Visualization and testing of team profiles
Höschl, Cyril
The aim of this thesis is to demonstrate methods for visualization and testing team profiles. An individual's profile is a vector of scores representing personality of a team member. Team profile is a collection of individual profiles. The first step for visualization of team profile is to find similarities among individual profiles. We present new approach for measuring similarity between two profiles as a significance of difference expressed by probability. We use this measure both for testing mean difference between two subgroups of profiles and also for building similarity matrix. We represent team profiles as a 3D ladscape map called Sociomap. To create a sociomap we demonstrate own algorithm for projecting similarity matrix into a 2D plane. We compare this algorithm with well known Multidimensional scaling and Principal Component Analysis methods solving the same task. Comparing with MDS and PCA, our algorithm provides more accurate result with comparable demand of time. We also employ some well known Computer Graphics methods to improve visual 3D representation of the sociomap and we also show some new improvements to render sociomap more realistic.
|
|
|
Advanced Moment-Based Methods for Image Analysis
Höschl, Cyril ; Flusser, Jan (advisor) ; Papakostas, George (referee) ; Jiřík, Radovan (referee)
The Thesis consists of an introduction and four papers that contribute to the research of image moments and moment invariants. The first two papers focus on rectangular decomposition algorithms that rapidly speed up the moment calculations. The other two papers present a design of new moment invariants. We present a comparative study of cutting edge methods for the decomposition of 2D binary images, including original implementations of all the methods. For 3D binary images, finding the optimal decomposition is an NP-complete problem, hence a polynomial-time heuristic needs to be developed. We propose a sub-optimal algorithm that outperforms other state of the art approximations. Additionally, we propose a new form of blur invariants that are derived by means of projection operators in a Fourier domain, which improves mainly the discrimination power of the features. Furthermore, we propose new moment-based features that are tolerant to additive Gaussian image noise and we show by extensive image retrieval experiments that the proposed features are robust and outperform other commonly used methods.
|
guest ::
