|
|
Connection of algorithms for removal of influence of skin diseases on the process for fingerprint recognition
Heidari, Mona ; Derawi, Mohammad (referee) ; Gomez-Barrero, Marta (referee) ; Drahanský, Martin (advisor)
Tato práce se zaměřuje na datové struktury, zpracování obrazu a metody počítačového vidění pro detekci a rozpoznávání nemocí ve snímcích otisků prstů. Počet vyvinutých biometrických systémů a dokonce i používaných biometrických charakteristik se zvyšuje. Všeobecně platí, že otisk prstu jednotlivce je jedinečný a zůstává relativně neměnný po celý život. Struktura papilárních linií se však může měnit nemocemi a může být poškozena kožními chorobami. Vzhledem k tomu, že jsou systémy do značné míry závislé na struktuře papilárních linií jednotlivce, která pozitivně ovlivňuje jejich identitu, lidé trpící kožními nemocemi mohou být diskriminováni, protože jejich papilární linie mohou být narušeny. Vliv kožních onemocnění je důležitým, ale často opomíjeným faktorem v biometrických systémech založených na otiscích prstů. Jedinec trpící kožním onemocněním, které postihuje konečky prstů nemusí být schopen používat určité biometrické systémy. Shromáždění databáze otisků prstů, ovlivněných kožními nemocemi, je náročný úkol. Je nákladný a časově náročný, vyžaduje také pomoc lékařských odborníků a ochotné účastníky trpící různými kožními nemocemi na bříšcíeh prstů. Surová databáze otisků prstů s onemocnénímí byla nejprve analyzována, aby poskytla pevný základ pro budoucí výzkum. Pro každé konkrétní onemocnění jsou nalezeny společné znaky mezi všemi snímky otisků prstů postižených nemocí a je definován obecný popis každého onemocnění a jeho vlivů. Poté automaticky přiřadíme označení na základě kombinace známého stavu obrazu otisku prstu. Navrhované řešení je integrováno s různými algoritmy zaměřenými na knihovny pro zpracování obrazu a metody počítačového vidění pro detekci objektů. Je vyhodnoceno na poškozených souborech dat otisků prstů a popisuje současný stav implementace pomocí navržených technik. Současný stav techniky pro implementaci detekce onemocnění využívá analýzu textury a detekci prvků porovnáváním hodnot intenzity pixelů v malém okolí v obraze. Vzhledem ke složitosti jednotlivých vzorů nemocí vede kombinace algoritmů analýzy textury k lepším výsledkům detekce. Kombinace Gray Level Co-occurrence Matrix (GLCM), Local Binary Pattern (LBP), pole orientací a matematické morfologie může detekovat poškození v obrazech otisků prstů. Kombinace těchto funkcí umožňuje identifikovat změny v textuře a tvaru toku papilárních linií otisků prstů způsobené nemocemi. Tyto techniky zachycují různé aspekty textury a tvaru poškození v obrazech otisků prstů a vedou k identifikaci změn v textuře způsobených nemocemi. V průběhu detekčního procesu jsou použity matematické morfologické operace pro zlepšení strukturálních detailů tím, že odstraňují malé nesrovnalosti v obraze a zjednodušují tvar objektů, což usnadňuje jejich identifikaci a izolaci, rozšiřováním hranic objektů v obraze nebo vyplněním mezer a propojením rozlomených částí objektů. To vede k lepší detekci a rozpoznání objektů.Na konci procesu detekce je použita koherence, která ukazuje hodnocení kvality polí obrazu otisku prstu na tři typy: zdravý, poškozený a pozadí.
|
|
|
The decision boundary
Gróf, Zoltán ; Hynčica, Tomáš (referee) ; Jirsík, Václav (advisor)
The main aim of this master's thesis is to describe the subject of the implementation of decision boundaries with the help of artificial neural networks. The objective is to present theoretical knowledge concerning this field and on practical examples prove these statements. The work contains basic theoretical description of the field of pattern recognition and the field of feature based representation of objects. A classificator working on the basis of Bayes decision is presented in this part, and other types of classificators are named as well. The work then deals with artificial neural networks in more detail; it contains a theoretical description of their function and their abilities in the creation of decision boundaries in the feature plane. Examples are shown from literature for the use of neural networks in corresponding problems. As part of this work, the program ANN-DeBC was created using Matlab, for the generation of practical results about the usage of feed-forward neural networks for the implementation of decision boundaries. The work contains a detailed description of this program, and the achieved results are presented and analyzed. It is shown as well, how artificial neural networks are creating decision boundaries in the form of geometrical shapes. The effects of the chosen topology of the neural network and the number of training samples on the success of the classification are observed, and the minimal values of these parameters are determined for the successful creation of decision boundaries at the individual examples. Furthermore, it's presented how the neural networks behave at the classification of realistically distributed training samples, and what methods can affect the shape of the created decision boundaries.
|
|
|
Controlling Computer Using Gestures
Lacko, Peter ; Herout, Adam (referee) ; Juránek, Roman (advisor)
This work deals with creation of system for controlling computer through webcam with gestures. Gesture in this work can be viewed as hand motion forming some pattern. In the beginning are described methods for hand detection, hand tracking and pattern recognition. Afterwards comes description of system and it's implementation with tests evaluation. Outcome of this work is program for simple control of document viewer and multimedia player.
|
|
|
Boosting and Evolution
Mrnuštík, Michal ; Juránek, Roman (referee) ; Hradiš, Michal (advisor)
This thesis introduces combination of the AdaBoost and the evolutionary algorithm. The evolutionary algorithm is used to find linear combination of Haar features. This linear combination creates the feature to train weak classifier for AdaBoost. There are described basics of classification, Haar features and the AdaBoost. Next there are basic information about evolutionary algorithms. Theoretical description of combination of the AdaBoost and the evolutionary algorithm is included too. Some implementation details are added too. Implementation is tested on the images as part of the system for face recognition. Results are compared with Haar features.
|
|
|
Similarity Search in Network Security Alerts
Štoffa, Imrich ; Kučera, Jan (referee) ; Žádník, Martin (advisor)
Network monitoring systems generate a high number of alerts reporting on anomalies and suspicious activity of IP addresses. From a huge number of alerts, only a small fraction is high priority and relevant from human evaluation. The rest is likely to be neglected. Assume that by analyzing large sums of these low priority alerts we can discover valuable information, namely, coordinated IP addresses and type of alerts likely to be correlated. This knowledge improves situational awareness in the field of network monitoring and reflects the requirement of security analysts. They need to have at their disposal proper tools for retrieving contextual information about events on the network, to make informed decisions. To validate the assumption new method is introduced to discover groups of coordinated IP addresses that exhibit temporal correlation in the arrival pattern of their events. The method is evaluated on real-world data from a sharing platform that accumulates 2.2 million alerts per day. The results show, that method indeed detected truly correlated groups of IP addresses.
|
|
|
Car Licence Plate Detection and Recognition
Kovaříček, Roman ; Procházka, Boris (referee) ; Váňa, Jan (advisor)
This bachelor thesis deals with finding the license plates in the image and pattern recognition. Work describes short history of the state license plates. It deals with also the current state license plates and their problems. It analyzes the process of image segmentation and follow evaluation of selected areas. A part of the work is design and implementation of algorithms that solve candidate search areas or characters. The final step is the recognition of individual characters and display the user with details of the result.
|
|
|
Objects Classification in Images
Gabriel, Petr ; Petyovský, Petr (referee) ; Janáková, Ilona (advisor)
This master's thesis deal with problems of classification objects on the basis of atributes get from images. This thesis pertain to a branch of computer vision. Describe possible instruments of classification (e.g. neural networks, decision tree, etc.). Essential part is description objects by means of atributes. They are imputs to classifier. Practical part of this thesis deal with classification of object collection, which can be usually found at home (e.g. scissors, compact disc, sticky, etc.). Analyzed image is preprocessed , segmented by thresholding in HSV color map. Then defects caused by a segmentation are reconstructed by morfological operations. After are determined atribute values, which are imputs to classifier. Classifier has form of decision tree.
|
|
|
Camera stereo-pair for object tracking in 3D
Manga, Marek ; Richter, Miloslav (referee) ; Babinec, Tomáš (advisor)
The first part deals with a design of simple pattern system that would allow tracking position of the object in space. There is described design of patterns and their properties are evaluated in several tests. The second part is devoted to a brief introduction on stereovision and calculation of the spatial coordinates of the object. Furthermore, there is described the resulting application for object tracking. Finally, is done repeabillity measure test.
|
|
|
Object Detection in Images
Vaľko, Tomáš ; Motlíček, Petr (referee) ; Švub, Miroslav (advisor)
Object detection in images is quite popular topic for years. What stands for it are a lot of works from this area of computer science. This thesis is about object classification, specifically human faces, which are one of the most interesting objects for processing. For classification we use neural networks, learned on face database. We study what influence has size of face database and preprocessing of digital image on neural network learning. This project implements simple face detector and localizator. It summarizes more and less successful results and indicates possible ways of system development in the future.
|
|
|
Machine learning for analysis of MR images of brain
Král, Jakub ; Říha, Ivo (referee) ; Provazník, Ivo (advisor)
The thesis is focused on methods of machine learning used for recognising the first stage of schizophrenia in images from nuclear-magnetic resonance. The introduction of this paper is focused primarily on physical principles. Further in this work, the attention is given to registration methods, reduction of data set and machine learning. In the classification part, simmilarity rates, support vectors´ method, K-nearest neighbour classification and K-means are described. The last stage of theoretical part is focused on evaluation of the clasification. In practical part the results of reduction data set by methods PCA, CRLS-PCA and subjects PCA are described. Furthermore, the practical part is focused on pattern recognition by methods K-NN, K-means and test K-NN method on real data. Abnormalities which are recognised by some classification methods can distinguish patients with schizophrenia from healthy controls.
|
guest ::
