|
|
Optimization of Basic Magnetic Field Homogeneity in MR Tomography
Hadinec, Michal ; Brančík, Lubomír (referee) ; Frollo, Ivan (referee) ; Bartušek, Karel (advisor)
This thesis is concerned with problems of measuring and mapping of magnetic field in MR tomograph, for purpose of magnetic flux density homogeneity optimization. Attention is paid to mapping techniques on rotary symmetric volume and to ways of magnetic fields optimization with utilization of passive and active correction systems. Theoretical analysis of magnetic field decomposition with utilization of spherical harmonics and numerical decomposition is made. Mapping and approximation techniques of basic magnetic field are verified by experiments in the laboratory at the Institute of Scientific Instruments AS CR in Brno.
|
|
|
Study of the influence of magnetic fields on biological tissue
Vlachová Hutová, Eliška ; Fiala, Pavel (referee) ; Bartušek, Karel (advisor)
The aim of this research is to explore the influence of magnetic field on early somatic embryos (ESEs) of spruce and pine. The magnetic field is gradual. This is important for the experiment, each cluster of ESE was exposed to magnetic field of different intensity. The resulting report of this experiment consists of the comparison between various methods of image analysis and the difference in growth of ESEs due to different exposure to magnetic field and surrounding temperature. Enclosed is a proposal of ESE size computing algorithm from a given image.
|
|
|
Susceptibility weighted magnetic resonance imaging (SWI)
Bačovský, Jaromír ; Bartušek, Karel (referee) ; Starčuk, Zenon (advisor)
Susceptibility weighted magnetic resonance imaging (SWI) seems to be a very promising method usable in the diagnosis of a number of clinical and biomedical applications. Difusse axonal injuries and neurodegenerative diseases to name a few. The goal of my bachelor's thesis was to study the principal methods of magnetic resonance imaging with particular emphasis on susceptibility weighted imaging. Then use this theoretical knowledge for practical implementation of the SWI method to 4,7T/200mm NMR system located at Institute of Scientific Instruments of the ASCR, v.v.i. It was necessary to create an image calculated software, which is capable of enhancing contrast caused by magnetic susceptibility differences as they manifest themselves in local phase changes between tissues. Few basic problems had to be solved. The phase image was filtered by a high pass filter, a Phase Mask was created from adjusted image, which is generated from phase unwrapped image. To varify the SWI method, samples were designed, which have the desired magnetic properties. During the initial phase of experimenting, hen's eggs were used and then citrus fruits with nanoiron, CaCl2, Zn and AgCl. We investigated basic parametres and settings of the MRI measurements that affect final susceptibility weighted image.
|
|
|
Precise segmentation of image data
Svoboda, Jan ; Marcoň, Petr (referee) ; Mikulka, Jan (advisor)
The concern of this thesis is a development of an extension module for 3D Slicer platform. The core of the module is an implementation of a Support Vector Machines classifier, which is used for segmentation of the vertebral column image data provided by the University Hospital Brno. One of the goals of the thesis was resampling and registration of these image sequences. CT volumes provided solid contrast and were used as a reference for gaining properly segmented groups of vertebrae. Due to the low quality of the MRI volumes image data, segmentation of MRI images was not completely succesful. The extension module scripted in Python language can be seen as a tool and can be used in the future for different datasets.
|
|
| |
|
|
Neuroinformatics: calibration methods in multicentric MR studies
Kovalčík, Tomáš ; Říha, Ivo (referee) ; Provazník, Ivo (advisor)
Work deals with methods of calibration of multi-center study of magnetic resonance imaging. Magnetic resonance is the phenomenon of the substance in a magnetic field of induction B0 delivers energy in the form of RF pulse with the Larmor frequency and thus to excite particles to higher energy levels. Calibration can be performed using the distinctive and homogeneous RF phantoms. Furthermore, we can perform calibration using image registration. To calibrate the images by registering the work described in the classical linear (affine) and nonlinear. Listed below are the simulators, which are also useful for modeling various artifacts.
|
|
|
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.
|
|
| |
|
|
Automatic 3D segmentation of brain images
Bafrnec, Matúš ; Dorazil, Jan (referee) ; Kolařík, Martin (advisor)
This bachelor thesis describes the design and implementation of the system for automatic 3D segmentation of a brain based on convolutional neural networks. The first part is dedicated to a brief history of neural networks and a theoretical description of the functionality of convolutional neural networks. It represents a fast introduction to the problematics and provides theoretical basics needed for the understanding and creation of the system. Individual layers of the neural network and principles of their functionality and mutual relations are also described in this part. The second part of the thesis is about problem analysis, designing of a solution and a comparison between neural networks and other solutions. The result of a magnetic resonance imaging of the head is a series of black-and-white images representing a 3D scan. The task is to tag a brain and to remove unnecessary information in the form of surrounding tissues. The final image of the brain can be utilized in a volumetry or during a diagnostic of neurodegenerative diseases. The advantage of neural networks in comparison with deterministic systems is their flexibility. They allow an adaptation to other segmentation problems just by changing the training dataset, without a need of changes in the architecture. One of the systems performing fully automatic 3D segmentation is called U-Net – its name comes from the similarity of the architecture with the letter U. Three real solutions, the first implementation of U-Net, extended U-Net and recurrent U-Net were presented. The first version of U-Net has been very memory-demanding, it required a training on a processor instead of a graphic card and has not allowed data processing in full resolution. The extended U-Net has resolved these problems by loading data in overlaying series of three images. In addition to the possibility of a training on a graphic card with related decrease in learning time, the accuracy was increased by adding interconnections to the internal architecture of the network. The last version, recurrent U-Net, aims for the optimization of extended U-Net based on the reusage of existing levels. This brings a decrease in a time and resource difficulty. The number of parameters of the network was lowered to less than 20%, without any increase in case of further level addition. This network is one of first recurrent networks used on the problem of 3D segmentation and provides a foundation to further research. The last part focuses on the evaluation of results and the comparison of accuracy, speed and requirements between particular networks. The accuracy of human and machine segmentation is also compared. The extended and recurrent U-Net have surpassed their human opponent, which in real case could save a lot of doctors time and prevent human mistakes. The result of this work is a theoretical basis providing an introduction to the problematics of convolutional neural networks and segmentation, fully working systems for automatic 3D segmentation and the foundation for further research in the field of recurrent networks.
|
|
|
Creating a knowledge base for the diagnosing of diseases
Macháček, Daniel ; Steinerová, Kateřina (referee) ; Jirsík, Václav (advisor)
This bachelor thesis is focused on problematic of creation knowledge base. It is describing basics of expert systems, their function and possible usage in modern world. In result of this thesis is knowlenge base in web aplication NPS able to diagnose diseases of hematology-oncology and that is proving possibility for use in real life. Knowledge base was created in cooperation with experts in the medical field and contains real data.
|
guest ::
