京都大学 / 新制・課程博士 / 博士(情報学) / 甲第23548号 / 情博第778号 / 新制||情||132(附属図書館) / 京都大学大学院情報学研究科通信情報システム専攻 / (主査)教授 湊 真一, 教授 山本 章博, 准教授 川原 純 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Observing the dynamic interaction between stars and their close stellar neighbors is key to establishing the stars’ orbits, masses, and other properties. Our ability to visually discriminate nearby stars is limited by the power of our telescopes, posing a challenge to astronomers at small observatories that contribute to binary star surveys. Masks placed at the telescope aperture promise to augment the resolving power of telescopes of all sizes, but many of these masks must be manually and repetitively reoriented about the optical axis to achieve their full benefits. This paper introduces a design concept for a mask rotation mechanism that can be adapted to telescopes of different types and proportions, focusing on an implementation for a Celestron C11 Schmidt–Cassegrain optical tube assembly. Mask concepts were first evaluated using diffraction simulation programs, later manufactured, and finally tested on close double stars using a C11. An electronic rotation mechanism was designed, produced, and evaluated. Results show that applying a properly shaped and oriented mask to a C11 enhances contrast in images of double star systems relative to images captured with the unmasked telescope, and they show that the rotation mechanism accurately and repeatably places masks at target orientations with minimal manual effort. Detail drawings of the mask rotation mechanism and code for the software interface are included.

Finite tree automata is formalism used in many different areas of computer science, among others in the area of formal verification. Nowdays there are few tools used for handling of finite tree automata, however libraries of MONA tool are the best choice. The finite tree automata are a frequent tool for formal verification of computer systems which work with dynamic data structures. The input format of finite tree automata for libraries of MONA tool is very difficult for humans because it is necessary to enter the move function of the finite tree automaton in a form of several multiterminal binary decision diagrams. The aim of this thesis is to design and implement tool to convert the finite set of move rules into internal format of the MONA tool.

The presented thesis deals with the issue of statistical quality control of a specific production process. The thesis presents a range of statistical tools that can be used to identify the factors causing a high proportion of non-standard pieces. The diploma thesis practically introduces the application of these quality management tools to the production process characterized by an increased proportion of non-standard pieces, in which the factors causing their occurrence are unknown, and as following the thesis in detail introduces the approach how to detect these factors. The last part of the work summarizes the recommendations handed over to the company in order to verify the conclusions of the thesis.

Performance of automatic speech recognition (ASR) systems utilizing only acoustic information degrades significantly in noisy environments such as a car cabins. Incorporating audio and visual information together can improve performance in these situations. This work proposes a lip detection and tracking algorithm to serve as a visual front end to an audio-visual automatic speech recognition (AVASR) system. Several color spaces are examined that are effective for segmenting lips from skin pixels. These color components and several features are used to characterize lips and to train cascaded lip detectors. Pre- and post-processing techniques are employed to maximize detector accuracy. The trained lip detector is incorporated into an adaptive mean-shift tracking algorithm for tracking lips in a car cabin environment. The resulting detector achieves 96.8% accuracy, and the tracker is shown to recover and adapt in scenarios where mean-shift alone fails.

This thesis is about using neural networks in recognition of letters A to Z and numbers 0 to 9. In the first part is theoretically described substance of neural networks and concretically described principle the method of learning multiple-layer network with backward spreaded error(a.ka Backpropagation). Basic problematic of processing the picture and resilence of network against degradation picture by a noise and compression JPEG is also described here. Second part is directed to practical realization of feed foward multiple-layer network with recognition the binary patterns of alphabetical letters and numbers 0 to 9, which was created in Matlab and Simulink environment. Next and final part is about practical realization of feed foward network with recognition the grayscale patterns of alphabetical letters and numbers 0 to 9, which was also created in Matlab and Simulink environment.

The aim of this diploma thesis is to explore methods of face detection and recognition in the picture. The method for face detection and the method for face recognition will be chosen according to literature survey. Both methods will be implemented using the OpenCV library and a program language C/C++. The result of this project is creation of graphic interface which use programmed function for face detection and recognition from a picture and also a camcorder.

Efficient manipulation of Boolean functions is an important component of many computer-aided design task. As a data structure for representing and manipulating Boolean functions, Binary Decision Diagrams are commonly used. These diagrams are commonly used in many fields such as model checking, system verification, circuit design, etc. In this thesis we describe these diagrams and there are present their modifications. Furthermore, this paper present and describes techniques for effective handling and representation of binary decision diagrams. This thesis describes the design and implementation of library that will work with these diagrams. It is further discussed how the developed library can be used within the library VATA for manipulating tree automata. Finally, the library was compared with well known and heavily optimized library CUDD, which is public and with library CacBDD. The experimental results showed that the performance of the proposed library is quite close to that of CUDD a CacBDD (has comparable and mostly even slightly better performance).

Image segmentation is an important step in image processing. A traditional way how to segment an image is a texture-based segmentation that uses texture features to describe image texture. In this work, Local Binary Patterns (LBP) are used for image texture representation. Texture feature is a histogram of occurences of LBP codes in a small image window. The work also aims to comparison of results of various modifications of Local Binary Patterns and their usability in the image segmentation which is done by unsupervised clustering of texture features. The Fuzzy C-Means algorithm is finally used for the clustering in this work.

Au cours des dernières années, les exigences croissantes en matière de dissipation thermique à haut rendement pour la microélectronique, les engins spatiaux, les réacteurs nucléaires, etc., encouragent le développement d'échangeurs de chaleur de nouvelle génération. Le caloduc est l’un des équipements de refroidissement efficaces et potentiels. La plupart du transfert de masse et de chaleur se fait au niveau de la micro-région près de la ligne triple de contact (solide, liquide, vapeur), qui est essentielle à l'amélioration de la performance thermique du caloduc. Cette étude se concentre sur le processus d'évaporation de gouttes sessiles de deux nouveaux fluides de travail (solution binaire et nanofluide), qui possèdent une micro-région similaire à celle du caloduc. Le flux de Marangoni induit par le gradient de concentration et la conductivité thermique exceptionnelle devraient améliorer significativement le débit evaporé du mélange alcool-eau et du nanofluide de graphène, respectivement. Une combinaison de techniques acoustiques et infrarouges est développée pour suivre la variation de la concentration d'alcool pendant l'évaporation des gouttes des mélanges 1-butanol-eau et éthanol-eau. Selon l'observation du comportement d'évaporation à différentes températures du substrat, une série d'équations empiriques est suggérée pour prédire le taux d'évaporation de la solution binaire de 1-butanol-eau en considérant l'effet Marangoni thermal et solutal. De plus, l'effet de la PEGylation, de la concentration des nanoparticules et de la température du substrat sur l'évaporation de gouttes de graphène nanofluide est étudié par des méthodes microscopiques, optiques et infrarouges. Les résultats expérimentaux et l'analyse thermodynamique peuvent contribuer à la compréhension complète du mécanisme impliqué concernant les performances d'évaporation du nanofluide de graphène. / In recent years, increasing requirement in high efficient heat dissipation for micro-electronics, spacecraft, nuclear reactors etc., encourage the development of next generation heat exchanger. Heat pipe is one of potential effective cooling equipments and most of mass and heat transfer take place at micro-region near triple phase (solid, liquid, vapor) contact line of working fluid, which is essential to thermal performance improvement of heat pipe. This study focuses on the evaporation process of sessile droplets of two novel working fluids (binary solution and nanofluid), which possess similar micro-region to that in heat pipe. Concentration gradient induced Marangoni flow and exceptional thermal conductivity are expected to significantly enhance evaporation rate of alcohol-water mixture and graphene nanofluid, respectively. A combination of acoustic and infrared techniques is developed to track alcohol concentration variation during evaporation of 1-butanol and ethanol aqueous droplets. According to observation of evaporation behavior at different substrate temperature, a series of empirical equations is suggested to predict evaporation rate of 1-butanol-water binary solution droplet considering thermal and solutal Marangoni effect. In addition, the effect of PEGylation, nanoparticle concentration and substrate temperature on drop evaporation of graphene nanofluid are investigated by microscopic, optical and infrared methods. Experimental results and thermodynamic analysis can contribute to the full understanding of involved mechanism concerning evaporation performance of graphene nanofluid.