|
|
Biometrics using Face Recognition
Koupil, Michal ; Mézl, Martin (referee) ; Odstrčilík, Jan (advisor)
This work is focused on a face detection in a picture and subsequent recognition of the face in its respective database. Face detection had been implemented using Viola-Jones algorithm. To recognize the face afterwards, PCA (Principal Component Analysis) a LBPH (Local Binary Pattern Histograms) had been used. Implemented algorithm had been tested on freely accesible biometric databases. Overall success rate of face detection was 93,4 %. Overall success rate of face recognition using PCA was 88,1 % and with LBPH it was 93,1 %. Of the two methods for face recognition, LBPH method has better ability to perform recognition with data, that do not possess ideal parameters for biometrics. With data suitable for biometrics, both of the methods perform well, with PCA being faster.
|
|
|
USB Ultrasound Imaging System
Kratochvíl, Tomáš ; Mézl, Martin (referee) ; Harabiš, Vratislav (advisor)
This master thesis deals with development and construction of ultrasound diagnostical system with USB interface. The main task is to study current usage of A mode ultrasound systems and methods of their construction. Following with own design of A mode ultrasound system with usage of assigned Olympus ultrasound transducer.
|
|
|
Expert System for Detection of Ventricular Extrasystoles
Svánovská, Zuzana ; Mézl, Martin (referee) ; Sekora, Jiří (advisor)
Ventricular premature beats are ectopic changes in the ECG signal. Detection of ventricular extrasystoles on 12leads ECG was created in MATLAB. At first my work involves whether the ventricular premature beats comes from the right or the left ventricle. Another algorithm evaluates the incidence outflow tract of the chamber. If the previous algorithm shows signs of indecision, another algorithm is used for detection of ventricular extrasystoles at another location in the left ventricle. The last step is to narrow the area of detection which makes it easier to find the place of origin.
|
|
|
Acquisition, Modeling and Signal Processing in Ultrasound Perfusion Imaging
Mézl, Martin ; Kozubek, Michal (referee) ; Flusser,, Jan (referee) ; Jiřík, Radovan (advisor)
This work deals with possibilities of ultrasound perfusion analysis for the absolute quantification of perfusion parameters. In the theoretical part of this work are discussed possibilities of using of the ultrasound contrast agents and approaches for the perfusion analysis. New methods for the perfusion analysis are suggested and tested in the practical part of this work. The methods are based on convolutional model in which the concentration of the contrast agent is modeled as aconvolution of the arterial input function and the tissue residual function. The feasibility of these methods for the absolute quantification of perfusion parameters is shown on data from phantom studies, simulations and also preclinical and clinical studies. The software for the whole process of the perfusion analysis was developed for using in hospitals.
|
|
|
Image segmentation of spinal disc in medical imaging
Meloun, Jan ; Nemček, Jakub (referee) ; Mézl, Martin (advisor)
This thesis is focused on segmentation of intervertebral discs in images from two medical imaging modalities - computed tomography (CT) and magnetic resonance imaging (MRI). Theoretical introduction of the thesis describes intervertebral disc herniation and relevant imaging modalities. It also includes description of basic and advanced segmentation methods. For practical part of the thesis, three different segmentation teqniques (one for CT data and two for MRI images) have been chosen, implemented and applied on images acquired at Radiodiagnostic department of Havlíčkův Brod hospital. The segmentation quality has been assessed quantitatively for individual methods by comparing the segmentation results to manually created reference segmentation of intervertebral discs.
|
|
|
Real-time ECG monitoring
Valentová, Vanessa ; Mézl, Martin (referee) ; Šikner, Tomáš (advisor)
The key element of ECG signal analysis is the detection of the QRS complex. The QRS complex is the most significant part of the signal. In this work, there is a short summary of methods used for QRS detection in real time. Based on previous research, the detection alghorithm is chosen and created in LabVIEW. The paper is also focused on cardiac arrhytmias. Created alghorithm is used for detecion certain types of arrhytmia. Alghorithm was tested on MIT/BIH arrhytmia database.
|
|
|
Finger vein biometrics
Bělehrádek, Stanislav ; Sekora, Jiří (referee) ; Mézl, Martin (advisor)
This thesis presents basic concepts in the field of biometrics, with a focus on optical methods of finger vein biometrics. The paper also contains additional information including physics and anatomy terms, which are important for understanding presented problems. Part of the thesis is draft of the device used for finger vein biometry and proposal of effective image preprocessing process for biometric purpose. Main part of the thesis contains creation and testing of such device. User interface is created in MATLAB and allows identification and verification of people with record in database. Biometric system is tested on group of 16 individuals. Achieved results are discussed and compared with results of similar devices.
|
|
|
Analysis of MR data
Guzmická, Lea ; Mézl, Martin (referee) ; Fabíková, Veronika (advisor)
The object of the work is diffusion MRI. The work deals with it from acquisition methods to postprocessing methods. The thesis is divided into 7 chapters. These gradually focus on the basic principles of MRI, measuring parameters, the phenomenon of diffusion and various methods of dMRI. The second part is comparing dMRI softwares and direct on actual postprocessing and analysis of data of study comparison of heathy controls and patients suffering eating disorders.
|
|
|
Classification of glioma grading in brain MRI
Olešová, Kristína ; Mézl, Martin (referee) ; Chmelík, Jiří (advisor)
This thesis deals with a classification of glioma grade in high and low aggressive tumours and overall survival prediction based on magnetic resonance imaging. Data used in this work is from BRATS challenge 2019 and each set contains information from 4 weighting sequences of MRI. Thesis is implemented in PYTHON programming language and Jupyter Notebooks environment. Software PyRadiomics is used for calculation of image features. Goal of this work is to determine best tumour region and weighting sequence for calculation of image features and consequently select set of features that are the best ones for classification of tumour grade and survival prediction. Part of thesis is dedicated to survival prediction using set of statistical tests, specifically Cox regression
|
|
|
Evolutionary algorithms for global optimization problem solving
Dragon, Ondřej ; Kozumplík, Jiří (referee) ; Mézl, Martin (advisor)
This work is devoded to evolutionary algorithms and solution of global optimization problems, mainly the traveling salesman problem. The traveling salesman problem is analyzed in detail as well as its methods of solution, such as: graph theory, heuristics and evolutionary algorithms. The main optimization method of this work is a Inver - over operator. In conclusion are implemented selected methods and performed testing and evaluation of the individual data sets.
|
guest ::
