|
|
Palm liveness detection and identification
Kala, Adam ; Goldmann, Tomáš (referee) ; Sakin, Martin (advisor)
This bachelor's thesis is focused on the detection of liveness of the human hand and its identification from multispectral images. The main goal of the work is the recognition of a living hand and its forgeries based on the information we have obtained, the finding of minutiae for palm identification and their use. The work describes techniques for cropping the area of interest, extracting characteristic features from the palm, and subsequent detection of liveness and identification of the person. Liveness detection success reached over 90% and person identification with a False Acceptance Rate (FAR) of less than 2% for selected datasets.
|
|
|
Detection of Material Surface Damage Based on a Photograph
Marek, Radek ; Sakin, Martin (referee) ; Dyk, Tomáš (advisor)
This work focuses on the use of various types of neural networks for detecting surface damage of materials from photographs and evaluates their effectiveness. Identifying different types of damage, such as cracks, scratches, and other defects, is essential for assessing the condition of materials and may indicate the need for further maintenance or repairs. The use of advanced neural networks allows for more precise detection and classification of damage, which is crucial for applications in areas such as construction, the automotive industry, and aerospace engineering, where rapid and reliable diagnostics of material defects are critical. Integrating these technologies into regular inspection processes can significantly improve accident prevention and extend the lifespan of structural components. The work also discusses the possibilities for improvement and adaptation of algorithms to specific materials and types of damage. Thus, this work demonstrates how advanced machine learning technologies can significantly contribute to more effective and reliable material condition monitoring, opening paths for future innovations in maintenance and safety.
|
|
|
Precise Reconstruction of Damaged Parts in Fingerprint Images
Šmotláková, Lucia Mária ; Sakin, Martin (referee) ; Kanich, Ondřej (advisor)
The aim of the bachelor's thesis is to locate the damaged areas in fingerprints and then reconstruct them using cubic Bézier curves. Structure elements and morphological operations are used to localize the damage. The break points obtained by localization are used to plot the curve. The outcome is a simple user interface that allows the user to upload a fingerprint image and observe each step of the solution of the bachelor's thesis. The achieved results were then tested and analyzed. In conclusion, it is found that the use of Bézier curves for reconstruction is not suitable for all types of damage, in particular it is not suitable for large damage in areas of high curvature.
|
|
|
Algorithm for Facial Image Quality Estimation
Husár, Tomáš ; Sakin, Martin (referee) ; Goldmann, Tomáš (advisor)
The precision of face recognition algorithms is heavily influenced by the quality of input images. The aim of the work is to evaluate the quality of face images using a convolutional neural network. The data on which the testing was carried out were created by various degradations of photos from the CelebA dataset. The resulting application determines the quality of images based on the predicted probabilities of individual degradations.
|
|
|
Integration of tools working with fingerprints
Dobeš, Kristián ; Sakin, Martin (referee) ; Kanich, Ondřej (advisor)
This bachelor thesis deals with the design and implementation of the integration of tools working with fingerprints into current application SyFDaS. The main goal was to extend existing features of the application and develop new ones, such as visualization of simulated damage, damage annotation and generation of random input parameters for simulation. In the work, a methodology for integrating the new tools into the application was presented and successfully demonstrated. During the course of the work, the current user interface was greatly enhanced to better support different types of damage. The integration and extensions of tools were implemented in the C# programming language, including a reworking of the tools from Python using the EmguCV library for image processing. In the final part, quality analysis of the integrated tools with new extensions and testing of user interface friendliness were performed to confirm the functionality and usability of the application.
|
|
|
Animal Identification Based on Biometric Information
Jančeková, Lucia ; Sakin, Martin (referee) ; Dyk, Tomáš (advisor)
Táto práca sa zaoberá identifikáciou divej zvery, konkrétne diviaka pomocou fotografii nosu. Ide o identifikáciu jednotlivcov a využívajú sa na to ryhy, ktoré sa nachádzajú na vrchnej časti nosu. V rámci tejto práci je navrhnuté a implementované riešenie na extrakciu tejto biometrickej informácii a porovnaním ju s ostatnými už uloženými šablónami. Riešenie je otestované na fotografiách z rovnakého diviaka, ale aj na 49 ďalších individuálnych jedincov.
|
|
|
Liveness Detection of a Hand
Vanický, Jozef ; Kanich, Ondřej (referee) ; Sakin, Martin (advisor)
This thesis solves the security problem of biometric systems which uses biometric data of a hand. It focuses on the absence of liveness detection during verification of a sample by the biometric system. In result this thesis offers to extend the contactless biometric systems by a system of liveness detection, which utilizes specific properties of human skin and its reactions to certain wavelengths of the visible spectrum of electromagnetic radiation. Reactions of the skin on the specific wavelengths of visible spectrum are analyzed and compared with the reactions of inorganic materials which imitates the hand. Outcome of this thesis is a design and implementation of a liveness detection algorithm, which can, based on the analysis of a set of photographs of the human palm, which was exposed to four wavelengths of a visible part of the electromagnetic spectrum, successfully recognize a hand from a counterfeit with an average accuracy rate of 82.20 %.
|
|
|
Extraction of Detailed Information from Plastic Fingerprints
Krsička, Tomáš ; Sakin, Martin (referee) ; Kanich, Ondřej (advisor)
This thesis deals with automatic analysis of plastic fingerprints and extraction of statistic information from them. The application uses a texture highlighting algorithm to separate the area that includes the fingerprint and extracts the width and density of papillary ridges. These values are then converted into mm and appropriately exported. This information can then be used to estimate the age of the owner of the fingerprint. The results of the analysis of most of the data have shown wider ridges in male fingerprints and a growing width of ridges with the age of the owner.
|
|
|
Morphing of Fingerprints
Dovičic, Denis ; Sakin, Martin (referee) ; Kanich, Ondřej (advisor)
This bachelor`s thesis deals with the creation of a realistic looking fingerprint with a double identity. The goal is the design and subsequent implementation of an algorithm to create such a morphed fingerprint. The morphing consists of three important steps, extracting the information from the fingerprint, morphing the information to generate a new fingerprint, and verifying the morphed fingerprint against both originals to evaluate the double identity rating. As an extension, an algorithm for blurring the edges of a morphed fingerprint was implemented, the Minutia Cylinder-Code SDK was integrated into the application to immediately obtain double identity rating, and the method of obtaining suitable identities from the original fingerprints was modified to improve results. The implemented algorithms were tested on a fingerprint database, where it was shown that in the absence of damage, the information from the fingerprint is reliably extracted, from 90 % cases the morphed fingerprint acquired two identities and from 60 % it also had a realistic appearance.
|
|
|
Extraction of Features from the Palm for Identification of Persons
Pricl, Patrik ; Kanich, Ondřej (referee) ; Sakin, Martin (advisor)
The work deals with the identification of persons using the palm. The specific properties of the skin and blood vessels are used. The response of the skin and blood vessels to electromagnetic radiation is used to capture these properties. Gabor's filter is used to extract skin properties. The LLBP algorithm is used to extract the veins. It uses the Hamming distance, the Euclidean distance and the K-nearest neighbor algorithm to compare properties.
|
guest ::
