National Repository of Grey Literature 49 records found  1 - 10nextend  jump to record: Search took 0.01 seconds. 
Automatic Image Analysis for Production Quality Control of Textile
Sýkorová, Tereza ; Dobeš, Petr (referee) ; Zemčík, Pavel (advisor)
This work deals with the classification of defects that occur in the production of nonwovens. The defect classification task is part of a system for automatic production quality control. The goal is to implement a method that will classify problematic defect classes with sufficient accuracy. That was achieved using convolutional neural networks (CNN). The best results were achieved by the EfficientNet network, which had an accuracy of 81% when evaluated by cross-validation on an available dataset. Within the work, a number of experiments are performed, which are focused on the modification of input data. The influence of the shape and composition of the input images on the final classification is examined. A CNN model was also implemented, which uses additional information for classification in addition to the image.
Generative Neural Networks for Handwritten Text
Ševčík, Pavel ; Dobeš, Petr (referee) ; Hradiš, Michal (advisor)
The aim of this study was to create a generative neural network for handwritten text lines. The model produces variable-sized images of handwritten text lines based on the expected style. The proposed method exceeds existing models in the image quality and can be used to generate both individual words and entire lines of handwritten text. It combines the use of the attention mechanism to extract the features for each character from the text query and their arranging on the line by inserting spaces between them. The new approach allows more granular control of the symbol positions on the line, which leads to smoother style interpolations. In contrast to the previous approach, the proposed method uses the Gaussian filter to spread the individual symbols features to the surrounding area. This approach also allows to train the model for symbols position predictions using the adversarial loss (GAN). In addition, annotations of symbol horizontal positions on the lines of the IAM dataset of handwritten text have been created.
Consequences of NAFTA Trade Agreement for the Car Industry in North America
Dobeš, Petr ; Kozák, Kryštof (advisor) ; Fiřtová, Magdalena (referee)
The subject of this thesis is the North American Free Trade Agreement (NAFTA), signed between the United States of America, Canada and Mexico and its impact on the automotive industry in North America between the years 1994, when NAFTA came into force, and 2009, when two major American car manufacturers, General Motors and Chrysler, went bankrupt during the global-scale recession and the industry changed significantly. The thesis is based on the theory of comparative advantages, as it was described by a British economist David Ricardo in the 19th century. It subscribes to the general principle that a free trade is beneficial to all engaged parties, because it enables more effective allocation of resources and provides for more specialization of production. The thesis argues NAFTA was a complex and ambitious international trade deal that had profound impact on the evolution of this branch of industry in the United States, Canada and Mexico, however its impact on the economy as a whole was limited and many of the changes, attributed to NAFTA, would likely have happened even without its passage due to the natural process of evolution of the industry and modernization. The creation of a continent-wide zone of free trade enabled local and foreign car makers to establish international supply chains that...
Fluid inclusions in gold-bearing quartz gangue from Padrť and Sobětice localities
Hemalová, Kateřina ; Zachariáš, Jiří (advisor) ; Dobeš, Petr (referee)
Quartz veins with molybdenite and gold from the locality Padrť crosscut metamorphosed Cambrian and Ordovician sediments (quartzite, arcose, cherts) of the Barrandien unit in southwest part of Central Brdy Mts.. Based on fluid inclusion microthermometry we distinguish three main generations of the quartz gangue: Q1 - the oldest quartz, that forms the main portion of gangue; Q2 - xenomorphic crystals growing on Q1, subdivided into Q2a (dark nuclei of crystals with a quantity of primary fluid inclusions) and Q2b (pellucid crystals crystallized to vugs) overgrowing Q2a; Q3 - the youngest quartz (with calcedony-like texture) that overgrowths Q2b crystals. The first generation of quartz (Q1) precipitated from low salinity (~5 wt. % eq. NaCl) aqueous-carbonic fluid with minor methane/nitrogen admixture (~ 5 mol. %). Estimated PT conditions of Q1 formation are >350 řC and ~ 400-500 MPa (depth about 15 km under lithostatic pressure). Formation of quartz Q2 and Q3 is associated with aqueous fluids. Q2 precipitated from low salinity (< 5,9 wt. % eq. NaCl) fluid at 250 to 320 řC and 60 to 120 MPa (depth about 3 - 5 km under hydrostatic pressure). Younger subgeneration Q2b contains quantity of fluid iinclusions with signs of boiling and with wide range of salinity 1,2 to 7 wt. % eq. NaCl. The Q3 was formed from even...
Interaction of proteins with inhibitors: quantum chemical study
Dobeš, Petr ; Hobza, Pavel (advisor) ; Vondrášek, Jiří (referee) ; Jurečka, Petr (referee)
This dissertation focuses on theoretical studies of the interaction between protein kinases and their inhibitors. Studied protein kinases, cyclin-dependent kinase 2 (CDK2) and CK2 kinase (casein kinase 2) play an important role in regulating cellular processes in eukaryotic organisms. Their abnormal function in human cells can lead to serious diseases. This process can be stopped by blocking the aberrant protein kinases using specific low molecular weight inhibitors. Inhibitors of protein kinases typically bind to the active site of the enzyme by noncovalent interactions. Theoretical description of these interactions using quantum-chemical and molecular mechanical methods can help in understanding the biophysical principles governing the binding. These, in turn, can be subsequently used for a rational drug design of more effective and more specific inhibitors. The stabilization energy of the complex of CDK2 with inhibitor roscovitine is predominantly formed by the dispersion energy. DFT methods, which do not describe the dispersion energy was thus completely inappropriate for the treatment of such a system. When an empirical term is included to correct for the description of dispersion, such methods, as e.g. the SCC-DFTB-D, can be recommended for computation of this or similar complexes. The dominant part...
Principles of application of fluid inclusions for the study of hydrothermal mineralizations
Štrba, Martin ; Zachariáš, Jiří (advisor) ; Dobeš, Petr (referee)
The bachelor thesis is subdivided into two review parts and one experimental part. The first part focuses on petrography and microthermometric study of fluid inclusions. Petrography chapter lists genetic types of inclusions, and various mechanisms that determine their shape and digree of fill. Emphasis is placed especially on the microthermometric study of fluid inclusions. The phase changes in H2O - NaCl and H2O - CO2 - salts systems at low and high temperatures are described in detail. A separate chapter is dedicated to interpretation of microthermometric data and to their use in determination of pressure-temperature conditions of minerals formation. The second part of the thesis lists several examples of using the fluid inclusions in the study of hydrothermal deposits and in prospecting of ore deposits (Pb - Zn - fluorite deposits of Mississippi valley type, porphy - type, Cu - ores at Bingham). One chapter is dedicated to Bleïda deposit (Morocco), Kanmantoo deposit (Australia) and San Cristobal vein (Peru). This chapter discribes geology, mineralogy and fluid inclusion data of these deposits. Third, experimental, chapter deals with hydrothermal mineralization in the Ševětín quarry. It includes the regional geology of Ševětín massif, geology of the quarry and the mineralogy of hydrothermal...
OCR Trained with Unanotated Data
Buchal, Petr ; Dobeš, Petr (referee) ; Hradiš, Michal (advisor)
The creation of a high-quality optical character recognition system (OCR) requires a large amount of labeled data. Obtaining, or in other words creating, such a quantity of labeled data is a costly process. This thesis focuses on several methods which efficiently use unlabeled data for the training of an OCR neural network. The proposed methods fall into the category of self-training algorithms. The general approach of all proposed methods can be summarized as follows. Firstly, the seed model is trained on a limited amount of labeled data. Then, the seed model in combination with the language model is used for producing pseudo-labels for unlabeled data. Machine-labeled data are then combined with the training data used for the creation of the seed model and they are used again for the creation of the target model. The successfulness of individual methods is measured on the handwritten ICFHR 2014 Bentham dataset. Experiments were conducted on two datasets which represented different degrees of labeled data availability. The best model trained on the smaller dataset achieved 3.70 CER [%], which is a relative improvement of 42 % in comparison with the seed model, and the best model trained on the bigger dataset achieved 1.90 CER [%], which is a relative improvement of 26 % in comparison with the seed model. This thesis shows that the proposed methods can be efficiently used to improve the OCR error rate by means of unlabeled data.

National Repository of Grey Literature : 49 records found   1 - 10nextend  jump to record:
Interested in being notified about new results for this query?
Subscribe to the RSS feed.