|
|
Modern coding of speech signals using overcomplete models
Zapletal, Ondřej ; Průša, Zdeněk (referee) ; Rajmic, Pavel (advisor)
The theoretical contents of this thesis are studies of overcomplete models. Those are the models of signals, on which is set for their parametrization more variables, than it's necessary and consequently there's computed so-called sparse solution via iteration algorithms. A goal of this analysis is a selection just of the considerable (sparse) parameters. The theory is based on a linear algebra, vector spaces, bases and so-called frames. The task of the individual project of this thesis is a description and simulation of two speech coders: a classical coder based on linear predictive speech coding and a coder, that's making use of overcomplete stochastic ARMA processes models. A part of their realization is to simulate their decoders and a analyze their reconstruction quality. For their realization there is used MATLAB and an overcomplete models' library (toolbox frames).
|
|
|
Eye tracking
Magula, Filip ; Rajmic, Pavel (referee) ; Vlach, Jan (advisor)
This work deals with tracing the eyes in real-time video using infrared radiation treatment and detection of its rebound on pupils. It describes the human eye and its response to infrared radiation. It contains proposals for hardware and software solutions and the final realization of the basic options in Simulink, explains the selected methods of digital image processing and includes an assessment of the methods used in this work in conclusion.
|
|
|
Real-Time Generation of Band-Limited Digital Audio Signals
Maule, Petr ; Rajmic, Pavel (referee) ; Schimmel, Jiří (advisor)
Master’s thesis deals with the generation of digital audio signals with band-limited frequency spectrum, i.e. without the aliasing distortion. Various methods of generating band-limited rectangular, triangular, and sawtooth waveforms are described in the theoretical part. The described methods are programmed in the Matlab programming environment and compared in terms of real-time parameter changes, such as duty cycle change of rectangular waveform or continuous change of frequency. The main part of the thesis describes implementation of methods of successive integration of band-limited impulse train and method of differentiated parabolic waveforms in C++ language. The implemented methods were integrated into a plug-in of VST technology that generates an audio signal in real time. The implemented methods are compared in terms of computational complexity and distortion of the generated signal.
|
|
|
Segmented wavelet transform of images
Kučera, Michal ; Rajmic, Pavel (referee) ; Průša, Zdeněk (advisor)
This master’s thesis is dealing with modification of classical discrete time wavelet transform algorithm to be able to split input image into several segments which could be independently processed. Segmented discrete time wavelet transform is introduced and implemented and the tests shows that it produces the similar values of wavelet coefficients as the classical approach. Complete independent image segment processing is allowed thanks to inverse segmented discrete time wavelet transform. It’s obvious that redundancy calculations appear when the image is processed segment by segment compared to transform the image at once. The redundancy rises with increasing decomposition depth and filter length, however it falls when the segment dimensions are increased. It happens because it is worked with smaller amount of segments which needs to extend. The extension with neighbor samples is the only source of redundancy. The thesis also contains description of two different approaches to signal border extension: firstly complete extension for all decomposition levels before the transformation itself and secondly classical method when the signal is extended in each decomposition step. Both approaches have the same results of image reconstruction, but it distinguishes in calculated wavelet coefficients. Four kinds of signal border extension are compared by using MSE and PSNR. The results of all approaches are similar, however the smooth padding of order 0 and symmetric-padding could be slightly favored against smooth padding of order 1 and zero-padding which shows higher MSE.
|
|
|
Software module generating "atoms" for purposes of overcomplete signal representation
Špiřík, Jan ; Průša, Zdeněk (referee) ; Rajmic, Pavel (advisor)
The aim of this master thesis is generating new "atoms'' for purposes of overcomplete signal representation for toolbox Frames in MATLAB. At first is described the principle of overcomplete systems and so-called frames. In the thesis is introduced the basic distribution of frames and conditions of their constructions. There is described the basic principle of finding the sparse solutions in overcomplete systems too. The main part is dealt with construction single functions for generating "atoms'', such as: Gabor function, B-splines, Bézier curves, Daubechies wavelets, etc. At last there is introduced an example of usage these functions for reconstruction signal in comparison with Fourier and wavelet transforms.
|
|
|
Objective methods for evaluation of the audio signal quality
Granát, Michal ; Rajmic, Pavel (referee) ; Rášo, Ondřej (advisor)
This diploma thesis deals with the objective methods of audio-signal qualification. In detail have been described algorithms to the reader which are used in individual methods. Individual methods are analyzed from computational intensity, ability to predict precisely subjective quality of audio-signals, point of view. There have been described basic variables by which audio-signals are described in further parts of this thesis. Second part has been dedicated to theoretical description of individual methods – their advantages and disadvantages and practical forms of usage. Third part has been dedicated to implementation of two chosen methods in graphical interface of MATLAB program. All the functions are described in detail which these applications offer. In the last part of the thesis there are in detail described results obtained by practical testing of audio-signals with different character in created applications and evaluated quality of reached results as an outcome.
|
|
|
Wavelet transform based object detector
Mikuš, Ondřej ; Průša, Zdeněk (referee) ; Rajmic, Pavel (advisor)
This thesis deals with applying methods on object detection in image. Separation of objects off the background is often needed during the image processing. It isolates the region of interest that can be worked with. The main purpose of this paper is the explanation of principles of pre-processing and segmentation of image, resulting in object detection using the wavelet transformation. This wavelet transformation is described more in detail, because it is the base of the primary used method. In the practical part of this thesis the main method was implemented to MATLAB environment and tested on set of images. The method was tested for robustness against noise and blur of image. It was compared with commonly used methods, using the edge detectors and thresholding. A simulation program was created for comparison of methods efficiency, including user interface.
|
|
|
Reverb Digital audio effect based on convolution with impulse response of acoustic room
Tichý, Vladimír ; Rajmic, Pavel (referee) ; Mačák, Jaromír (advisor)
This work deals with a computer simulation of an acoustic room using its impulse response. Two different approaches to the simulation are described with their pros and cons and then the work is focused on the physical approach, which uses room’s impulse response during the simulation. Several methods for the extraction of the impulse response of the acoustic room are mentioned with their conditions of use. The detailed description of various algorithms for a real time convolution computing is followed by the cost analysis of frequency domain block convolution algorithms. Several algorithms are chosen, implemented and tested in Matlab environment. Then the most effective of them is chosen to be implemented in VST technology as the plug in module for real time room simulation.
|
|
|
Graph theory - implementation of selected problems
Stráník, František ; Rajmic, Pavel (referee) ; Koutný, Martin (advisor)
This work is intended on identification with basic problems from the graphs theory area. There are the basic conceptions as well more complicated problems described. The one part of this work is specialized in working of individual types of graphs. It starts with single linked list through double linked list after as much as trees which represented the simplest graphs textures. The other part of this work devotes to the whole graph and describes more complicated problems and their resolution from the theory graphs area. Among these problems belongs to searching in graphs help by Depth First Search and Breadth First Search methods. Then searching the shortest way help by the specific algorithms as are: Dijkstra´s algorithm, Floyd-Warshall´s algorithm and Bellman-Ford´s algorithm. The last part is devoted to problems with searching minimal frames of graphs with usage Kruskal´s algorithm, Jarnik´s algorithm and Boruvka´s algorithm methods.
|
|
|
Speech denoising based on wavelet transform and voice recognition in segments
Chrápek, Tomáš ; Sysel, Petr (referee) ; Rajmic, Pavel (advisor)
The wavelet transform is a modern signal processing tool. The wavelet transform earned itself a great success mainly for its unique properties, such as the capability of recognizing very fast changes in processed signal. The theoretical part of this work is introduction to wavelet theory, more specifically wavelet types, a wavelet transform and its application in systems dealing with signal denoising. A main problem connected to speech signals denoising was introduced. The problem is degradation of the speech signal when denoising unvoiced parts. It is because of the fact that unvoiced parts and noise itself have very similar characteristics. The solution would be to apply different attitude to voiced and unvoiced segments of the speech. The main goal of this diploma thesis was to create an application implementing the speech signal denoising using the wavelet transform. The special attention should have been paid to applying different attitude to voiced and unvoiced segments of the speech. The demanded application is programmed as a grafical user interface (GUI) in MATLAB environment. The algorithm implemented in this form allows users to test introduced procedures with a great comfort. This work presents achieved results and discusses them considering general requirements posed on an application of given type. The most important conlusion of this Diploma Thesis is the fact that some kind of trade-off between sufficient signal denoising and keeping the speech understandable has to be made.
|
guest ::
