Refine
Year of publication
Document Type
- Bachelor Thesis (100)
- Diploma Thesis (81)
- Study Thesis (76)
- Master's Thesis (46)
- Doctoral Thesis (18)
- Part of Periodical (12)
- Conference Proceedings (3)
Language
- German (282)
- English (52)
- Multiple languages (2)
Keywords
- Bildverarbeitung (16)
- Augmented Reality (13)
- Robotik (10)
- Computergrafik (9)
- Computergraphik (9)
- Computervisualistik (9)
- OpenGL (8)
- GPGPU (5)
- GPU (5)
- Line Space (5)
Institute
- Institut für Computervisualistik (336) (remove)
live video performances at cultural events since 2005. Until now, we have been developing single video-performance applications that mostly met the demands of a certain project. What we did not achieve so far is a) a modular software design and, b) a feature for deskewing planar surfaces that are projected at a non perpendicular angle (assuming the use of visual projectors). This paper deals with solving the first problem by desiging a modular framework and the second problem by implementing an user-friendly module for deskewing planar surfaces. The deskewing process is completely manual, letting the user edit the coordinates of the surfaces" cornerpoints by dragging the points with the mouse. For this, the cornerpoints" x- and y-values are manipulated and the z-value is left constant. While in this way, the 3D-interaction with a 2D-interface can be avoided, the exclusivley two-dimensional transformation of planar surfaces introduces undesired texture-mapping artifacts produced by the triangulation-based rendering of graphic-cards. In order to avoid these artifacts, a selected method called "adaptive subdivision" is presented that corrects the introduced rendering errors.
Simulation of fractures
(2014)
Real-time computing often avoids the simulation of fractures due to its complexity. The field of engineering science provides methods to create these simulations to improve games and other applications. Steadily rising computer capacities allow suitable simulations on a real-time basis and make this aspect increasingly interesting. The topic and aim of this research is to simulate fractures of stiff bodies. The primary objective is the physical plausibility and performance of the application. This thesis analyses the potential of computer science to realize the simulation of fractures.
Three existing as well as one self-created were implemented and analysed. The works "Real time simulation of deformation and Fracture of stiff material" from Müller et al., "real time simulation of Brittle Fracture using Modal analysis" from Glondu et al. and "Fast and Controllable simulation of the Shattering of Brittle Objects" from Smith et al. form the basis of this thesis. The introduced methods use different computation of forces and fractures. The developed procedure uses the idea of generating secondary breaks. The approaches were implemented based on the Bullet physics-engine. The results of the work show that physically based breaks are realizable on a real-time basis.
The analysis of the physical methods demonstrates that their performance mainly depends on the constitution of the used objects. This thesis shows that the further investigation of this topic can discover new possibilities. The improvement of the realism in virtual worlds can be achieved by executing physically plausible methods.
In this bachelor thesis a system for the simulation of the movements of molecules is developed. The calculation of the forces between chemically bonded atoms as well as intermolecular forces is done almost entirely on the GPU. The visualization of the simulation happens at an interactive framerate. To achieve rendering in realtime on off-the-shelf graphics cards, apt optimizations and slight abstractions of the underlying physical models are needed. One can control the execution speed or completely stop the simulation at any given moment. Some of the parameters of the underlying physical models of the simulation can be modified at runtime. With the right settings for the parameters, some phenomena of molecular dynamics can be observed, for example the spacial structure of the molecules.
The animation of models has become an important part in different areas of everyday life. It is a demanding task for computer graphics to generate
a natural deformation of organic models. Skinning is a common method to animate models without animating each vertex individually. The skin of the model deforms automatically by manipulating individual bones of a skeleton.
This bachelor thesis deals with the most common algorithm, linear blend skinning and aims to find some optimizations regarding the visual effect
and performance. Additionally it presents certain instancing methods which are combined with the skinning methods in the application to show the advantages and disadvantages of the latter.
In no other field of computer science has the hardware been evolved more
quickly than in computer graphics. Therefore the GPU offers, aside from
the pure rendering of triangles, a bunch of further pipeline steps that allows
visualisation of other graphics objects, like freeform surfaces.
This bachelor’s thesis is about the rendering of freeform surfaces, in particular
bezier surfaces. For that reason an implementation for management
and visualisation of bézier surfaces was created for the rendering framework
of the university Koblenz (CVK). For this purpose first a triangulation
was implemented and finally a tesselation of bezier surfaces with normals
and texture coordinates, as well as the handling of trim curves.
The present work starts with an introduction of methods for three-dimensional curve skeletonization. Different kinds of historic and recent skeletonization approaches are analysed in detail. Later on, a state-of-the-art skeletonization algorithm is introduced. This algorithm deals as a basis for the own approach presented subsequently. After the description and definition of a new method improving the state-of-the-art algorithm, experiments are conducted to get appraisable results. Next, a ground truth is described which has been set up manually by humans. The human similarity evaluations are compared with the results of the automatic computer-based similarity measures provided by the own approach. For this comparison, standard evaluation criteria from the field of information retrieval have been used.
Die Studienarbeit analysiert mit Hilfe einer erweiterten Balancetheorie die Relationen des Roboters zu den Personen im Kamerabild und den Personen untereinander. Es wurde gezeigt, dass die Abstraktion der Balancetheorie auf eine konkrete Anwendung übertragen werden kann. Allerdings muss die Theorie erweitert und teilweise eingeschränkt werden, um sie flexibler, aber gleichzeitig passend zur Anwendung zu gestalten. Dadurch wird die theoretische Grundlage, auf der die Arbeit beruht, ebenfalls verändert. Damit diese Modifikationen nicht den Rahmen der psychologischen Grundlage verlassen, müssen sie wiederum mit psychologischen Mitteln untersucht werden. Das würde allerdings den Umfang dieser Arbeit überschreiten, aber bietet ein Ansatz für eine interdisziplinäre Zusammenarbeit der Psychologie und Robotik. Die Interaktion und das Verhalten der Maschinen nach menschlicher Vorlage zu gestalten, ist für beide Disziplinen von Interesse. Im Kontext der Entwicklung einer ausreichenden Interaktion zwischen der Maschine und dem Menschen, wäre es interessant zu erforschen, welche sozialen Merkmale detektiert werden müssten, zum Beispiel im Spektrum der Mimik. Darüberhinaus ermöglicht die Recognize-Pipeline ein Ansatz, die Interaktion über das Kamerabild hinaus zu verfolgen. Dadurch kann der Roboter das Gesicht mit vergangenen Interaktionen assoziieren und dem entsprechend agieren. Allerdings bedarf die Pipeline-Struktur weiterer Arbeit. So werden bestehende Datenbanken über ein Gesicht nicht mit neuen Bildern erweitert, so fern sie notwendig sind. Auch kann keine automatische Korrektur erfolgen, falls fehlerhafte Informationen in die Datenbank gelangen. So kann es vorkommen, dass das selbe Gesicht zwei unterschiedliche IDs erhält, wenn das Gesicht nicht wiedererkannt wird. Auch können sehr ähnliche Gesichter zusammenfallen zu einer ID. Solche Fehler müssten für eine stabile Anwendung selbständig korrigierbar sein.
Ist es möglich, allein mittels VR-Headset bei Nutzern Immersion zu
erzeugen? Zur Beantwortung dieser Frage werden zwei Simulationen einer
Achterbahnfahrt ohne haptisches Feedback mittels der Unreal Engine
4.20.3 für ein HTC-Vive VR Headset entwickelt und implementiert. Die
zweite Simulation unterscheidet sich von der ersten durch die Darbietung
außergewöhnlicher Ereignisse während der Fahrt, für die vermutet wird,
dass sie das Immersionserleben verstärken. Elf Probanden nahmen an der
Untersuchung teil. Die Auswertung eines Fragebogens zur Erfassung der
Intensität der Immersion und der Antworten auf offenen Fragen zeigt, dass
Immersion in beiden Simulation erfolgreich erzeugt werden konnte. Manche
Merkmale der Simulation vertieften bei einzelnen Probanden das immersive
Erleben, bei anderen dagegen nicht. Die Bedeutung der Ergebnisse
und Optimierungsmöglichkeiten für künftige Studien werden diskutiert.
In dieser Arbeit werden mehrere Verfahren zur Superresolution, die zwei unterschiedlichen Ansätzen zuzuordnen sind, implementiert und miteinander verglichen. IBP, ein Verfahren, welches der Tomographie ähnelt, stellt den klassischen Ansatz der Superresolution dar. Das Prinzip von IBP ist leicht verständlich und relativ einfach zu implementieren, hat aber den Nachteil, keine eindeutigen Lösungen zu produzieren, da es schwer ist, Vorwissen mit einfließen zu lassen. Bei den statistischen Verfahren erweist es sich als äußerst zeitkritisch, die Systemmatrix M vorzuberechnen. Capel und Zisserman haben beim Maximum-Likelihood-Verfahren eine starke Rauschanfälligkeit festgestellt. Der Vergleich ergibt, dass IBP in punkto Bildqualität Maßstäbe setzt, zumindest bei kaum verrauschten Eingangsbildern und guter Registrierung. Allerdings stellen viele Eingangsbilder wegen des linearen Laufzeitverhaltens ein Problem dar. Die statistischen Verfahren liefern bei stark verrauschten Eingangsbildern bessere Ergebnisse als IBP. Durch viele Eingangsbilder lassen sich diese noch weiter verbessern. Die bei vorberechneten Systemmatrizen nahezu konstante Laufzeit und ihre Robustheit bei Rauschen prädestinieren die statistischen Verfahren für solche Bildserien. Ungenau registrierte Eingangsbilder wiederum führen bei allen Verfahren zu unscharfen Ergebnissen.
Ziel dieser Diplomarbeit ist es, katadioptrische Stereosysteme aufzubauen und zu testen. Hierfür gilt es, die Komponenten auszuwählen und zu beschaffen. Es soll ein System mit einem Planspiegel und eines mit einem gekrümmten Spiegel aufgebaut werden. Die Vorgehensweise in beiden Fällen unterscheidet sich jedoch deutlich voneinander. Das Planspiegel-System wird unter den Vorraussetzungen angefertigt, dass es möglichst kostengünstig ist und bietet sich vor allem deshalb für eine solches Vorgehen an, da es mit handelsüblichen Komponenten realisierbar ist. Die Verwendung dieser Komponenten bringt allerdings Einschränkungen in der Bildqualität mit sich. Hinzu kommt noch die Tatsache, dass es sich um eine Computational Camera mit geringer Komplexität handelt, da es direkt rektifizierte Ergebnisse liefert und aufgrund dessen leicht zu implementieren ist. Diese Kombination von Eigenschaften macht es zu einem geeigneten Kandidaten für eine solche Herangehensweise. Das radiale System wird hingegen unter der Prämisse maximaler Qualität gebaut. Der verwendete Spiegel ist eine Einzelanfertigung und zusätzlich wird ein zirkuläres Fisheye-Objektiv benötigt, was die Kosten für ein solches System relativ hoch werden lässt. Des Weiteren ist die Dekodierung der Rohaufnahmen deutlich komplexer, als bei den bereits rektifizierten Aufnahmen des Planspiegel-Systems. Nach erfolgreichem Aufbau der beiden Gesamtsysteme sollen diese unter qualitativen Aspekten miteinander verglichen werden.
This examination broaches the issue of an investigation that is supported by the eye-tracking technologie and uses the e-learn-system Compass. For this investigation, there was carried out an experiment to test participants if they recognize agreement errors in the subjekt-verb- congruence in german sentences. The idea for that came from an other survey, that was hold for the englisch language. The integration of the e-learn-system Compass intended the observation of the participants learning behaviour. The experiment was able to be evaluated and analysed by the software of the eye-tracker, that recorded all eye movements and created valuable data out of them. It was assumed to catch the participants eye movements such as fixations or regressions.
The result was very clear: errors in congruence betweet subject and verb were done often. Besides, the error-rate increases proportionally to the length of the sentence.
In der vorliegenden Studienarbeit wird eine OpenGL-Applikation vorgestellt, die Geometrie-Shader in einem Feedback-Loop einsetzt, um auf der GPU Geometrie zu erzeugen. Dargelegt werden die erforderlichen Grundlagen Geometrie-Shader und Transform Feedback betreffend, die Umsetzung der Anwendung und die eingesetzten GLSL-Shader.
Der Schwerpunkt dieser Arbeit soll auf der schnellen sowie einfachen Umsetzung eigener Ideen von AR-Anwendungen liegen. Damit ein gewisser zeitlicher Rahmen bei der Umsetzung nicht überschritten wird, wurden Lösungen, die eine große Einarbeitungszeit oder fundierte Kenntnisse einer oder mehrer Programmiersprachen erfordern, nicht genauer betrachtet. Unter einer einfachen Umsetzung ist nicht zuletzt auch zu verstehen, dass diese auch dem Kreis der nicht-professionellen Anwender möglich sein soll. Dies beinhaltet, dass das gesuchte Programm auf normalen dem durchschnittlichen derzeitigen Stand der Technik entsprechenden Computersystemen lauffähig sein sollte. Auch wurden kommerzielle Produkte außer Acht gelassen, da die oft nicht unerheblichen Kosten einer nicht-professionellen Nutzung im Wege stehen.
Colonoscopy is one of the best methods for screening colon cancer. As the automatic detection of polyps in endoscopic images is a challenging task for image processing, a variety of research groups have proposed methods that try to fulfill this task to develop a system which supports the doctors during examination. However, the problem is still "at least partially" not solved. This paper gives a summary of 16 different polyp detection methods published in the last ten years. We found out that the major draw-back of many approaches is the lack of representative video data, which hinders comparison and evaluation of the published methods.
Efficient Cochlear Implant (CI) surgery requires prior knowledge of the cochlea’s size and its characteristics. This information helps to select suitable implants for different patients. Registered and fused images helps doctors by providing more informative image that takes advantages of different modalities. The cochlea’s small size and complex structure, in addition to the different resolutions and head positions during imaging, reveals a big challenge for the automated registration of the different image modalities. To obtain an automatic measurement of the cochlea length and the volume size, a segmentation method of cochlea medical images is needed. The goal of this dissertation is to introduce new practical and automatic algorithms for the human cochlea multi-modal 3D image registration, fusion, segmentation and analysis. Two novel methods for automatic cochlea image registration (ACIR) and automatic cochlea analysis (ACA) are introduced. The proposed methods crop the input images to the cochlea part and then align the cropped images to obtain the optimal transformation. After that, this transformation is used to align the original images. ACIR and ACA use Mattes mutual information as similarity metric, the adaptive stochastic gradient descent (ASGD) or the stochastic limited memory Broyden–Fletcher–Goldfarb–Shanno (s-LBFGS) optimizer to estimate the parameters of 3D rigid transform. The second stage of nonrigid registration estimates B-spline coefficients that are used in an atlas-model-based segmentation to extract cochlea scalae and the relative measurements of the input image. The image which has segmentation is aligned to the input image to obtain the non-rigid transformation. After that the segmentation of the first image, in addition to point-models are transformed to the input image. The detailed transformed segmentation provides the scala volume size. Using the transformed point-models, the A-value, the central scala lengths, the lateral and the organ of corti scala tympani lengths are computed. The methods have been tested using clinical 3D images of total 67 patients: from Germany (41 patients) and Egypt (26 patients). The atients are of different ages and gender. The number of images used in the experiments is 217, which are multi-modal 3D clinical images from CT, CBCT, and MRI scanners. The proposed methods are compared to the state of the arts ptimizers related medical image registration methods e.g. fast adaptive stochastic gradient descent (FASGD) and efficient preconditioned tochastic gradient descent (EPSGD). The comparison used the root mean squared distance (RMSE) between the ground truth landmarks and the resulted landmarks. The landmarks are located manually by two experts to represent the round window and the top of the cochlea. After obtaining the transformation using ACIR, the landmarks of the moving image are transformed using the resulted transformation and RMSE of the transformed landmarks, and at the same time the fixed image landmarks are computed. I also used the active length of the cochlea implant electrodes to compute the error aroused by the image artifact, and I found out an error ranged from 0.5 mm to 1.12 mm. ACIR method’s RMSE average was 0.36 mm with a standard deviation (SD) of 0.17 mm. The total time average required for registration of an image pair using ACIR was 4.62 seconds with SD of 1.19 seconds. All experiments are repeated 3 times for justifications. Comparing the RMSE of ACIR2017 and ACIR2020 using paired T-test shows no significant difference (p-value = 0.17). The total RMSE average of ACA method was 0.61 mm with a SD of 0.22 mm. The total time average required for analysing an image was 5.21 seconds with SD of 0.93 seconds. The statistical tests show that there is no difference between the results from automatic A-value method and the manual A-value method (p-value = 0.42). There is no difference also between length’s measurements of the left and the right ear sides (p-value > 0.16). Comparing the results from German and Egypt dataset shows there is no difference when using manual or automatic A-value methods (p-value > 0.20). However, there is a significant difference when using ACA2000 method between the German and the Egyptian results (p-value < 0.001). The average time to obtain the segmentation and all measurements was 5.21 second per image. The cochlea scala tympani volume size ranged from 38.98 mm3 to 57.67 mm3 . The combined scala media and scala vestibuli volume size ranged from 34.98 mm 3 to 49.3 mm 3 . The overall volume size of the cochlea should range from 73.96 mm 3 to 106.97 mm 3 . The lateral wall length of scala tympani ranged from 42.93 mm to 47.19 mm. The organ-of-Corti length of scala tympani ranged from 31.11 mm to 34.08 mm. Using the A-value method, the lateral length of scala tympani ranged from 36.69 mm to 45.91 mm. The organ-of-Corti length of scala tympani ranged from 29.12 mm to 39.05 mm. The length from ACA2020 method can be visualised and has a well-defined endpoints. The ACA2020 method works on different modalities and different images despite the noise level or the resolution. In the other hand, the A-value method works neither on MRI nor noisy images. Hence, ACA2020 method may provide more reliable and accurate measurement than the A-value method. The source-code and the datasets are made publicly available to help reproduction and validation of my result.
Human action recognition from a video has received growing attention in computer vision and has made significant progress in recent years. Action recognition is described as a requirement to decide which human actions appear in videos. The difficulties involved in distinguishing human actions are due to the high complexity of human behaviors as well as appearance variation, motion pattern variation, occlusions, etc. Many applications use human action recognition on captured video from cameras, resulting in video surveillance systems, health monitoring, human-computer interaction, and robotics. Action recognition based on RGB-D data has increasingly drawn more attention to it in recent years. RGB-D data contain color (Red, Green, and Blue (RGB)) and depth data that represent the distance from the sensor to every pixel in the object (object point). The main problem that this thesis deals with is how to automate the classification of specific human activities/actions through RGB-D data. The classification process of these activities utilizes a spatial and temporal structure of actions. Therefore, the goal of this work is to develop algorithms that can distinguish these activities by recognizing low-level and high-level activities of interest from one another. These algorithms are developed by introducing new features and methods using RGB-D data to enhance the detection and recognition of human activities. In this thesis, the most popular state-of-the-art techniques are reviewed, presented, and evaluated. From the literature review, these techniques are categorized into hand-crafted features and deep learning-based approaches. The proposed new action recognition framework is based on these two categories that are approved in this work by embedding novel methods for human action recognition. These methods are based on features extracted from RGB-D data that are
evaluated using machine learning techniques. The presented work of this thesis improves human action recognition in two distinct parts. The first part focuses on improving current successful hand-crafted approaches. It contributes into two significant areas of state-of-the-art: Execute the existing feature detectors, and classify the human action in the 3D spatio-temporal domains by testing a new combination of different feature representations. The contributions of this part are tested based on machine learning techniques that include unsupervised and supervised learning to evaluate this suitability for the task of human action recognition. A k-means clustering represents the unsupervised learning technique, while the supervised learning technique is represented by: Support Vector Machine, Random Forest, K-Nearest Neighbor, Naive Bayes, and Artificial Neural Networks classifiers. The second part focuses on studying the current deep-learning-based approach and how to use it with RGB-D data for the human action recognition task. As the first step of each contribution, an input video is analyzed as a sequence of frames. Then, pre-processing steps are applied to the video frames, like filtering and smoothing methods to remove the noisy data from each frame. Afterward, different motion detection and feature representation methods are used to extract features presented in each frame. The extracted features
are represented by local features, global features, and feature combination besides deep learning methods, e.g., Convolutional Neural Networks. The feature combination achieves an excellent accuracy performance that outperforms other methods on the same RGB-D datasets. All the results from the proposed methods in this thesis are evaluated based on publicly available datasets, which illustrate that using spatiotemporal features can improve the recognition accuracy. The competitive experimental results are achieved overall. In particular, the proposed methods can be better applied to the test set compared to the state-of-the-art methods using the RGB-D datasets.
Markerloses Tracking unter Verwendung von Analyse durch Synthese auf Basis von Featuredetektoren
(2008)
In der vorliegenden Diplomarbeit wurde ein auf "Analyse durch Synthese" sowie Featuredetektoren basierendes Trackingsystem implementiert, beschrieben und getestet. Das Ziel war die Untersuchung im Hinblick auf den Mehrwert der Computergraphik in einem markerlosen Trackingablauf, indem der Ansatz der "Analyse durch Synthese" zur Poseschätzung eingesetzt wird.