Szerző szerinti böngészés "Moghadasi, Mohammad"

Megjelenítve 1 - 2 (Összesen 2)
  • Bélyegkép
    TételKorlátozottan hozzáférhető
    Image processing
    (2012-06-01T08:37:58Z) Moghadasi, Mohammad; Fazekas, Gábor; DE--TEK--Informatikai Kar
    Biometrics-based authentication systems that use physiological and/or behavioral traits (e.g., fingerprint, face, and signature) are good alternatives to traditional methods. In spite of these advantages of biometric systems over traditional systems, there are many unresolved issues associated with the former. For example, how secure are biometric systems against attacks? How can we guarantee the integrity of biometric templates? How can we use biometric components in traditional access control frameworks? How can we combine cryptography with biometrics to increase overall system security? In this thesis, we address these issues and develop techniques to eliminate associated problems. Firstly, we analyze structure of biometric-based authentication systems and vulnerability of these systems and develop a method for increasing the security of image-based (e.g., fingerprint and face) biometric templates. For many years fractals were used for image compression. In the last few years they have also been used for object recognition. Fractal codes have been used as features for recognition, but we need to take into account the distance between codes, and to ensure the continuity of the parameters of the code. We use a method based on fractal image coding for face and fingerprint recognition. Generally, there are two approaches to using fractal image coding for recognition. The first type uses the fractal code itself, and discrimination comes from differences between the codes. The fractal code of an image is the parameters of a Partitioned Iterated Function System code generated for that image. The second type uses the decoding process of fractal image coding to perform recognition. It is more difficult to use the first type for recognition because fractal codes can change dramatically even between very similar images. The problem is that more than one fractal code can be generated for a given image. An advantage of using the second approach over the first is that we do not have to worry about the uniqueness and distance between codes. We only require the uniqueness of the attractor, which is already an implied property of a properly generated fractal code. ix This thesis focuses on the second approach, from which we develop a distance measure. Compared to other methods, research into fractal image coding based recognition methods has been scarce. Other advantages of this method are: increasing security because use of fractal codes instead of original images, compare fractal codes after decoding process and needless to compare geometrical parameters and non execution heavy preprocessing operation. Error rate in fingerprint verification is 7.1% and results in face verification in spite of different face expression are very acceptable.
  • Bélyegkép
    TételSzabadon hozzáférhető
    Multiple sclerosis Lesion Detection via Machine Learning Algorithm based on Converting 3D to 2D MRI Images by Using Value of Binary Pattern Classification and Computational Methods
    (2021) Moghadasi, Mohammad; Fazekas, Gábor; Informatikai tudományok doktori iskola; DE--Informatikai Kar -- DE--Informatikai Kar
    In the twenty first century, there have been various scientific discoveries which have helped in addressing some of the fundamental health issues. Specifically, the discovery of machines which are able to assess the internal conditions of individuals, has been a significant boost in the medical field. Multiple Sclerosis (MS) is a demyelinating disease in which the insulating covers of nerve cells in the brain and spinal cord are damaged. This damage disrupts the ability of the nervous system’s parts in various forms to transmit signals, therefore results a wide range of symptoms including physical, mental and psychiatric problems. In this regard, most of the researchers have focused on improving the classification of brain lesions, especially those from MS, which is a relatively difficult task, using different algorithms, but mainly in determining the edge. However, the absence of learning methods have caused complexity and difficulties for having accurate results in the previous works. Therefore, this need motivated us to use various tools of learning methods with a focus on Cellular Learning Automata (CLA) to achieve more accurate results in detection of MS lesions. Cellular Learning Automata (CLA) is a hybrid model of two, Learning Automata and Cellular Automata, which is a simple discrete system that can exhibit complex calculations and behavior through simple and local rules. In this study, we aim to propose a new combinational algorithm using Support Vector Machine (SVM) used for classification and cellular learning automata (CLA) to increase the accuracy of MS lesion detection. The objective is to create artificial models using support vector machines (SVM) to classify MS and normal brain MRI images, analyze the effectiveness of these models and their potential to use them in multiple sclerosis (MS) diagnosis. In order to develop such combination method, we start with simple learning methods such as k-means to find MS lesion. The research was carried out in four stages, respectively, the algorithms are as follows; a) Semi-automatic method and use of K-Means, b) Automatic MS Segmentation Approach Based on Cellular Learning Automata, c) MS Segmentation Approach based on SVM, CLA and K-Means, d) Accurate Simulated Database, 3D MRI to 2D Images, using value of Binary Pattern Classification for MS Detection. The algorithms consist of pre-processing parts, detecting MS-hemispheres, feature extraction, classification and post-processing. In the pre-processing section, the brightness intensity of the normalized images and the brain region are first extracted. Then, to reduce the computational volume, the lesions are diagnosed. The proposed approach can be considered as a supplementary or superior method for other methods such as Graph Cuts (GC), fuzzy c-means, mean-shift, k-Nearest Neighbor (KNN). We try to see the benefits of having a 3D database but to use 2D vectors only for better comparison and more accurate results. Support vector machines (SVM) can be a useful tool during the multiple sclerosis (MS) disease diagnosis process, however, to be able to make better assumptions, more tests are needed.