<div> Planar homography estimation is foundational to many computer vision problems, such as Simultaneous Localization and Mapping (SLAM) and Augmented Reality (AR). However, conditions of high variance confound even the state-of-the-art algorithms. In this report, we analyze the performance of two recently published methods using Convolutional Neural Networks (CNNs) that are meant to replace the more traditional feature-matching based approaches to the estimation of homography. Our evaluation of the CNN based methods focuses particularly on measuring the performance under conditions of significant noise, illumination shift, and occlusion. We also measure the benefits of training CNNs to varying degrees of noise. Additionally, we compare the effect of using color images instead of grayscale images for inputs to CNNs. Finally, we compare the results against baseline feature-matching based homography estimation methods using SIFT, SURF, and ORB. We find that CNNs can be trained to be more robust against noise, but at a small cost to accuracy in the noiseless case. Additionally, CNNs perform significantly better in conditions of extreme variance than their feature-matching based counterparts. With regard to color inputs, we conclude that with no change in the CNN architecture to take advantage of the additional information in the color planes, the difference in performance using color inputs or grayscale inputs is negligible. About the CNNs trained with noise-corrupted inputs, we show that training a CNN to a specific magnitude of noise leads to a ``Goldilocks Zone'' with regard to the noise levels where that CNN performs best.</div>

In first part thesis contains known methods of license plate detection. Preprocessing-based methods, AdaBoost-based methods and extremal region detection methods are described.Finally, there is a described and implemented own access using local detectors to creating visual vocabulary, which is used to plate recognition. All measurements are summarized on the end.

Early life stages of many marine and diadromous fish species use sandy beach surf zones, which occur along >50% of the world's marine coastlines. This extensive habitat can provide juvenile fishes with an abundant supply of potential prey and the ability to hide from predators in its shallow turbid waters. Chinook salmon is an anadromous species that migrates to the ocean during their first (subyearlings) or second (yearlings) year of life. The majority of subyearlings reside in estuaries during their first summer season; however, a small number of juveniles also use surf zones. Early marine residence is considered a critical period for Chinook salmon due to high mortality rates; however the role of surf zones in Chinook salmon life history is unclear. Therefore, I determined the distribution of juvenile Chinook salmon on beaches of the eastern North Pacific, compared the migration and growth patterns observed in surf zones and estuaries, identified the factors that accounted for variation in juvenile surf zone catch, explored the factors that influence growth rate variation in surf zones and estuaries, and modeled how growth rates in these coastal habitats may vary in the near future with predicted changes in climate. The majority (94%) of juveniles were caught in surf zones adjacent to estuaries with trough areas, which are beach sections where sand moved by currents and waves produce a trench-like shape. Surf zone fish were collected in significantly lower numbers than estuarine juveniles but entered brackish/ocean waters at similar sizes. Juveniles in surf zones consumed similar organisms (gammarid amphipods, crustacean larvae and insects) as in estuaries. Furthermore, stomach fullness indices (average = 2% of body weight) and growth rates (average = 0.4 mm day�����) were similar in surf zones and estuaries. At one surf zone, juvenile catch was positively correlated to short-term specific growth rates (14 days prior to capture). A bioenergetics modeling approach indicated that given current conditions, consumption rates accounted for more of the variation in growth than prey energetic content and temperature. Climate models predict future increases in fresh water temperature (1.5 to 5.8��C), sea surface temperature (1.2��C) and wave height (0.75 m) that could influence estuarine and surf zone use. Therefore, I developed a local mixing model based on these predictions to estimate future surf zone and estuarine water temperatures in two of the watersheds studied. Based on these temperature projections and the bioenergetics model, I predicted how juvenile specific growth rates would vary in both habitats. I determined that increases in water temperature in both habitats would reduce specific growth rates by 9 to 40% in surf zones and estuaries if diet composition and consumption rates remain similar to present conditions. To compensate for the decline in growth, juveniles may increase their consumption rates or consume more energetically rich prey, if available. If they are not able to compensate, their size at the end of the season may be reduced, which could reduce their overall survival. These results confirm that a small number of suyearling Chinook salmon use sandy beach surf zones, mostly adjacent to estuary mouths, where they experience growth conditions comparable to estuaries. My findings indicate that, in certain situations, juvenile Chinook salmon surf zone use can be influenced by surf zone growth conditions, while variation in growth rates are themselves most strongly influenced by variation in consumption rates in surf zones and estuaries. Predicted changes in coastal western North American climate will likely modify juvenile growth conditions in the next 50 years, and potentially reduce overall survival. Additional insights into the potential impacts of climate change on juvenile salmon will require estimates of changes in the composition, energetic quality and abundance of prey communities inhabiting coastal environments. / Graduation date: 2013

Les dernières années, la taille des collections vidéo a connu une forte augmentation. La recherche et la navigation efficaces dans des telles collections demande une indexation avec des termes pertinents, ce qui nous amène au sujet de cette thèse, l'indexation sémantique des vidéos. Dans ce contexte, le modèle Sac de Mots (BoW), utilisant souvent des caractéristiques SIFT ou SURF, donne de bons résultats sur les images statiques. Notre première contribution est d'améliorer les résultats des descripteurs SIFT/SURF BoW sur les vidéos en pré-traitant les vidéos avec un modèle de rétine humaine, ce qui rend les descripteurs SIFT/SURF BoW plus robustes aux dégradations vidéo et qui leurs donne une sensitivité à l'information spatio-temporelle. Notre deuxième contribution est un ensemble de descripteurs BoW basés sur les trajectoires. Ceux-ci apportent une information de mouvement et contribuent vers une description plus riche des vidéos. Notre troisième contribution, motivée par la disponibilité de descripteurs complémentaires, est une fusion tardive qui détermine automatiquement comment combiner un grand ensemble de descripteurs et améliore significativement la précision moyenne des concepts détectés. Toutes ces approches sont validées sur les bases vidéo du challenge TRECVid, dont le but est la détection de concepts sémantiques visuels dans un contenu multimédia très riche et non contrôlé.

Panoramas or stitched images are used in topographical mapping, panoramic 3D reconstruction, deep space exploration image processing, medical image processing, multimedia broadcasting, system automation, photography and other numerous fields. Generating real-time panoramic images in small embedded computer is of particular importance being lighter, smaller and mobile imaging system. Moreover, this type of lightweight panoramic imaging system is used for different types of industrial or home inspection. A real-time handheld panorama imaging system is developed using embedded real-time Linux as software module and Gumstix Overo and PandaBoard ES as hardware module. The proposed algorithm takes 62.6602 milliseconds to generate a panorama frame from three images using a homography matrix. Hence, the proposed algorithm is capable of generating panorama video with 15.95909365 frames per second. However, the algorithm is capable to be much speedier with more optimal homography matrix. During the development, Ångström Linux and Ubuntu Linux are used as the operating system with Gumstix Overo and PandaBoard ES respectively. The real-time kernel patch is used to configure the non-real-time Linux distribution for real-time operation. The serial communication software tools C-Kermit, Minicom are used for terminal emulation between development computer and small embedded computer. The software framework of the system consist UVC driver, V4L/V4L2 API, OpenCV API, FFMPEG API, GStreamer, x264, Cmake, Make software packages. The software framework of the system also consist stitching algorithm that has been adopted from available stitching methods with necessary modification. Our proposed stitching process automatically finds out motion model of the Spherical camera system and saves the matrix in a look file. The extracted homography matrix is then read from look file and used to generate real-time panorama image. The developed system generates real-time 180° view panorama image from a spherical camera system. Beside, a test environment is also developed to experiment calibration and real-time stitching with different image parameters. It is able to take images with different resolutions as input and produce high quality real-time panorama image. The QT framework is used to develop a multifunctional standalone software that has functions for displaying real-time process algorithm performance in real-time through data visualization, camera system calibration and other stitching options. The software runs both in Linux and Windows. Moreover, the system has been also realized as a prototype to develop a chimney inspection system for a local company. / Main Uddin-Al-Hasan, E-mail: main.hasan@gmail.com

This work focuses on a possibility of using current computer vision alghoritms and methods for automatic similarity matching of so called graffiti tags. Those are such graffiti, that are used as a fast and simple signature of their authors. The process of development and implementation of CBIR system, which is created for this task, is described. For the purposes of finding images similarity, local features are used, most notably self-similarity features.

This paper deals with object detection and tracking using iterest points. Existing approaches are described here. Inovated method based on Generalized Hough transform and iterative Hough-space searching is  proposed in this paper. Generality of proposed detector is shown in various types of objects. Object tracking is designed as frame by frame detection.

Úkolem této diplomové práce je navrhnout vlastní metodu pro zvýšení rozlišení v obraze scény, pokud je k dispozici více snímků dané scény. V teoretické části diplomové práce jsou jako nejlepší metody pro zvýšení rozlišení v obraze vybrány ty, které jsou založeny na principech zpracování signálu. Dále jsou popsány základní požadavky metod pro zvýšení rozlišení v obraze při přítomnosti více snímků stejné scény a jejich typická struktura. Následuje stručný přehled těchto metod a jejich vzájemné porovnání podle optimálních kritérií. Praktická část diplomové práce se zabývá samotným návrhem metody pro zvýšení rozlišení v obraze, pokud je k dispozici více snímků této scény. První navržená metoda je naimplementována a otestována. Při testování této metody je však  zjištěna její špatná funkčnost pro snímky scény s nízkým rozlišením, které vznikly vzájemnou rotací. Z toho důvodu je navržena vylepšená metoda pro zvýšení rozlišení v obraze. Tato metoda využívá při svém výpočtu robustních technik. Díky tomu je již vylepšená metoda nezávislá na rotaci mezi snímky scény s nízkým rozlišením. I tato metoda je řádně otestována a její výsledky jsou porovnány s výsledky první navržené metody pro zvýšení rozlišení v obraze. V porovnání výpočetních časů je lepší první navrhovaná metoda, avšak její výsledky pro obrazy obsahující rotace nejsou kvalitní. Oproti tomu pro obrazy, které vznikly pouze posunem při snímání scény, jsou tyto výsledky velice dobré. Vylepšená metoda je tedy využitelná zejména pro obrazy obsahující rotace. V závěru této práce je ještě navrženo jedno vylepšení, které by mohlo zlepšit výsledky druhé navrhnuté metody pro zvýšení rozlišení v obraze scény.

This work is focused on objects and scenes recognition using machine learning and computer vision tools. Before the solution of this problem has been studied basic phases of the machine learning concept and statistical models with accent on their division into discriminative and generative method. Further, the Bag-of-words method and its modification have been investigated and described. In the practical part of this work, the implementation of the Bag-of-words method with the SVM classifier was created in the Matlab environment and the model was tested on various sets of publicly available images.

This thesis deals with content based automatic photo categorization. The aim of the work is to experiment with advanced techniques of image represenatation and to create a classifier which is able to process large image dataset with sufficient accuracy and computation speed. A traditional solution based on using visual codebooks is enhanced by computing color features, soft assignment of visual words to extracted feature vectors, usage of image segmentation in process of visual codebook creation and dividing picture into cells. These cells are processed separately. Linear SVM classifier with explicit data embeding is used for its efficiency. Finally, results of experiments with above mentioned techniques of the image categorization are discussed.