|
|
Detecting a Partially Obscured Face in Image Data
Kedra, David ; Sakin, Martin (referee) ; Goldmann, Tomáš (advisor)
This bachelor thesis deals with the analysis of problems related to detecting partially occluded faces in camera systems and discusses current machine learning detectors. The aim was to find solutions useful for detecting hardly visible faces. For this reason, artificial occlusions were generated into datasets for training the YOLOv7, YOLOv7-tiny and RetinaNet models. A computer application that uses these detectors is presented. The models are compared with existing solutions in terms of quality and speed. The trained models perform better on most test datasets. YOLOv7 is the most accurate on the modified WIDER FACE and UFDD datasets with average precisions of 86 % and 89 % at a minimum IoU of 50 %. On the third dataset with face masks, the existing detector RetinaFace outperformes the others. According to the speed/quality ratio, YOLOv7-tiny is the most effective.
|
|
|
Forensic analysis of handwriting for the Czech environment using artificial intelligence
Stejskal, Jan ; Přinosil, Jiří (referee) ; Burget, Radim (advisor)
The analysis of handwriting is an important area of research in modern science. However, it is a very complex process because handwritten text can take on various forms. The use of artificial intelligence for analyzing and identifying text from different authors is nothing new in the world. Research in this area is, however, slightly lagging behind in the Czech environment. For this reason, several convolutional network architectures were proposed and compared in this work in an effort to find the most suitable structure for solving this problem. Of all the trained and tested models, the model based on the ResNet18 architecture achieved the highest accuracy, with a success rate of 92.2 % on a self-made database of 1328 samples with a resolution of 750x256. This result suggests that with a sufficiently large and high-quality database, the problem can be solved even in the Czech environment with its more complicated character set.
|
|
|
Learning the Face Behind a Voice
Zubalík, Petr ; Mošner, Ladislav (referee) ; Plchot, Oldřich (advisor)
The main goal of this thesis is to design and implement a system that will be able to generate a face based on the speech of a given person. This problem is solved using a system composed of three convolutional neural network models. The first one is based on the ResNet architecture and is used to extract features from speech recordings. The second model is a fully convolutional neural network which converts the extracted features into the styles which form a base for the final facial image. These styles are then passed as an input to the StyleGAN generator, which creates the resulting face. The proposed system is implemented in the Python programming language using the PyTorch framework. The last chapter of the thesis discusses some of the most significant experiments performed to fine-tune and test the developed system.
|
|
|
Classification of thorax diseases on chest X-ray images using artificial intelligence
Pijáček, Štěpán ; Mikulec, Marek (referee) ; Mezina, Anzhelika (advisor)
Tato práce se zabývá výzkumem použitelných řešení pro problém klasifikace onemocnění hrudníku na rentgenových snímcích hrudníku pomocí umělé inteligence. Pro lepší pochopení problému jsou v prvních kapitolách vysvětleny základní konvoluční neuronové sítě a jejich výhody a nevýhody. Na základě těchto prvních vysvětlení jsou vybrány dvě neuronové sítě, které rozšiřují koncept konvoluční neuronové sítě. Těmito sítěmi jsou kapslová síť a reziduální síť, obě jsou dále vysvětleny v příslušných kapitolách s jejich výhodami a nevýhodami. Reziduální síť a kapslová síť jsou poté implementovány pomocí programovacího jazyka python a frameworku TensorFlow s knihovnou Keras, obě se svými příslušnými kapitolami. Na konci práce jsou uvedeny výsledky a závěr.
|
|
|
Identification of supraventricular tachycardia segments using multiple-instance learning
Abbrent, Jakub ; Novotná, Petra (referee) ; Ronzhina, Marina (advisor)
Supraventricular tachycardias have a high incidence in the population and often cause health disorders. The aim of this thesis is to automatically detect and localize atrial fibrillation in ECG records. The algorithm, implemented in Python, uses a convolutional neural network ResNet for detection with multiple-instance learning and decision rules. The output of the detection in the form of a feature signal is used for localization. The classification achieved F1 score of 0.87 on the test dataset. Then, paroxysmal atrial fibrillation was localized with a deviation of -0.40±2.26 seconds for the onsets and 1.09±2.75 seconds for the terminations of the episodes. Lastly, the obtained results are evaluated and discussed.
|
|
|
Atrial Fibrillation Classification Using Deep Convolution Networks
Novotna, Petra
We propose the usage of three deep convolutional neural networks architectures for classification of a single lead electrocardiogram (ECG) recordings and evaluate them on the atrial fibrillation (AFIB) classification, for which data set was provided by the Department of Biomedical Engineering, BUT. The compared networks are based on ResNet, VGG net and AlexNet. Single lead signals are transformed into the form of spectrogram. AFIB data was augmented for the purpose of similar size of both respected classes and for successful classification. The most successful architecture, based on AlexNet, was found to perform obtaining an accuracy of 92 % and F1 score of 56 % on the hidden testing set.
|
|
|
Efficiency of deep convolutional neural networks on an elementary classification task
Prax, Jan ; Dobrovský, Ladislav (referee) ; Škrabánek, Pavel (advisor)
In this thesis deep convolutional neural networks models and feature descriptor models are compared. Feature descriptors are paired with suitable chosen classifier. These models are a part of machine learning therefore machine learning types are described in this thesis. Further these chosen models are described, and their basics and problems are explained. Hardware and software used for tests is listed and then test results and results summary is listed. Then comparison based on the validation accuracy and training time of these said models is done.
|
|
|
Self-supervised learning in computer vision applications
Vančo, Timotej ; Richter, Miloslav (referee) ; Janáková, Ilona (advisor)
The aim of the diploma thesis is to make research of the self-supervised learning in computer vision applications, then to choose a suitable test task with an extensive data set, apply self-supervised methods and evaluate. The theoretical part of the work is focused on the description of methods in computer vision, a detailed description of neural and convolution networks and an extensive explanation and division of self-supervised methods. Conclusion of the theoretical part is devoted to practical applications of the Self-supervised methods in practice. The practical part of the diploma thesis deals with the description of the creation of code for working with datasets and the application of the SSL methods Rotation, SimCLR, MoCo and BYOL in the role of classification and semantic segmentation. Each application of the method is explained in detail and evaluated for various parameters on the large STL10 dataset. Subsequently, the success of the methods is evaluated for different datasets and the limiting conditions in the classification task are named. The practical part concludes with the application of SSL methods for pre-training the encoder in the application of semantic segmentation with the Cityscapes dataset.
|
|
|
Graffiti Tags Re-Identification
Pavlica, Jan ; Beran, Vítězslav (referee) ; Špaňhel, Jakub (advisor)
This thesis focuses on the possibility of using current methods in the field of computer vision to re-identify graffiti tags. The work examines the possibility of using convolutional neural networks to re-identify graffiti tags, which are the most common type of graffiti. The work experimented with various models of convolutional neural networks, the most suitable of which was MobileNet using the triplet loss function, which managed to achieve a mAP of 36.02%.
|
|
|
Pedestrian Attribute Analysis
Studená, Zuzana ; Špaňhel, Jakub (referee) ; Hradiš, Michal (advisor)
This work deals with obtaining pedestrian information, which are captured by static, external cameras located in public, outdoor or indoor spaces. The aim is to obtain as much information as possible. Information such as gender, age and type of clothing, accessories, fashion style, or overall personality are obtained using using convolutional neural networks. One part of the work consists of creating a new dataset that captures pedestrians and includes information about the person's sex, age, and fashion style. Another part of the thesis is the design and implementation of convolutional neural networks, which classify the mentioned pedestrian characteristics. Neural networks evaluate pedestrian input images in PETA, FashionStyle14 and BUT Pedestrian Attributes datasets. Experiments performed over the PETA and FashionStyle datasets compare my results to various convolutional neural networks described in publications. Further experiments are shown on created BUT data set of pedestrian attributes.
|
guest ::
