Search result: Catalogue data in Spring Semester 2009
Computer Science Master  | ||||||
 Focused Courses | ||||||
| Focused Study: Visual Computing | ||||||
| Number | Title | Type | ECTS | Hours | Lecturers | |
|---|---|---|---|---|---|---|
| 251-0526-00L | Statistical Learning Theory | W | 5 credits | 2V + 1U | J. M. Buhmann | |
| Abstract | The course covers advanced methods of statistical learning : PAC learning and statistical learning theory;variational methods and optimization, e.g., maximum entropy techniques, information bottleneck, deterministic and simulated annealing; clustering for vectorial, histogram and relational data; model selection; graphical models. | |||||
| Objective | The course surveys recent methods of statistical learning. The fundamentals of machine learning as presented in the course "Introduction to Machine Learning" are expanded and in particular, the theory of statistical learning is discussed. | |||||
| Content | # Boosting: A state-of-the-art classification approach that is sometimes used as an alternative to SVMs in non-linear classification. # Theory of estimators: How can we measure the quality of a statistical estimator? We already discussed bias and variance of estimators very briefly, but the interesting part is yet to come. # Statistical learning theory: How can we measure the quality of a classifier? Can we give any guarantees for the prediction error? # Variational methods and optimization: We consider optimization approaches for problems where the optimizer is a probability distribution. Concepts we will discuss in this context include: * Maximum Entropy * Information Bottleneck * Deterministic Annealing # Clustering: The problem of sorting data into groups without using training samples. This requires a definition of ``similarity'' between data points and adequate optimization procedures. # Model selection: We have already discussed how to fit a model to a data set in ML I, which usually involved adjusting model parameters for a given type of model. Model selection refers to the question of how complex the chosen model should be. As we already know, simple and complex models both have advantages and drawbacks alike. # Reinforcement learning: The problem of learning through interaction with an environment which changes. To achieve optimal behavior, we have to base decisions not only on the current state of the environment, but also on how we expect it to develop in the future. | |||||
| Lecture notes | no script; transparencies of the lectures will be made available. | |||||
| Literature | Duda, Hart, Stork: Pattern Classification, Wiley Interscience, 2000. Hastie, Tibshirani, Friedman: The Elements of Statistical Learning, Springer, 2001. L. Devroye, L. Gyorfi, and G. Lugosi: A probabilistic theory of pattern recognition. Springer, New York, 1996 | |||||
| Prerequisites / Notice | Requirements: basic knowledge of statistics, interest in statistical methods. It is recommended that Introduction to Machine Learning (ML I) is taken first; but with a little extra effort Advanced Topics in Machine Learning can be followed without the introductory course. | |||||
| 251-0538-00L | Surface Representations and Geometric Modeling | W | 5 credits | 2V + 1U | M. Pauly | |
| Abstract | This course covers some of the latest developments in geometric modeling and surface representations. Topics include surface modeling based on triangle meshes, mesh generation, subdivision schemes, mesh fairing and simplification, multiresolution methods, and interactive shape editing. | |||||
| Objective | Introduction to geometric modeling and digital surface processing. The exercises will complement the course material with implementations of the main processing algorithms. | |||||
| Content | Recent advances in 3D digital geometry processing have created a plenitude of novel concepts for the mathematical representation and interactive manipulation of geometric models. This course covers some of the latest developments in geometric modeling and surface representations. Topics include surface modeling based on triangle meshes, mesh generation, subdivision schemes, mesh fairing and simplification, multiresolution methods, and interactive shape editing. | |||||
| Lecture notes | slides and course notes | |||||
| Prerequisites / Notice | Prerequisites: Introduction to Computer Graphics, experience with C++ programming. Some background in geometry or computational geometry is helpful, but not necessary. | |||||
| 251-0564-00L | Scientific Visualization | W | 5 credits | 2V + 1U | R. Peikert | |
| Abstract | Scientific visualization is the application of computer graphics to the visual analysis and interactive exploration of scientific data which have typically spatial or spatio-temporal domain. Such datasets arise in engineering, natural and medical sciences, and are generated by simulation, measurement or imaging techniques. | |||||
| Objective | Becoming familiar with the fundamental methods and some advanced techniques of scientific visualization. Being able to apply visualization to measurement or simulation data and to correctly interpret visualization results. | |||||
| Content | This course covers advanced topics in Scientific Visualization, including: contouring and isosurfaces, direct volume rendering, visualization of flow and vector fields, texture advection, feature extraction, topological methods, information visualization, visualization software, and hot topics of current research. | |||||
| 251-0570-00L | Game Programming Laboratory | W | 10 credits | 8P | B. Sumner, N. Thürey | |
| Abstract | The goal of this course is the in-depth understanding of the technology and programming underlying computer games. Students gradually design and develop a computer game in small groups and get acquainted with the art of game programming. | |||||
| Objective | The goal of this new course is to acquaint students with the technology and art of programming modern three-dimensional computer games. | |||||
| Content | This is a new course that addresses modern three-dimensional computer game technology. During the course, small groups of students will design and develop a computer game. Focus will be put on technical aspects of game development, such as rendering, cinematography, interaction, physics, animation, and AI. In addition, we will cultivate creative thinking for advanced gameplay and visual effects. The "laboratory" format involves a practical, hands-on approach with neither traditional lectures nor exercises. Instead, we will meet once a week to discuss technical issues and to track progress. We plan to utilize Microsoft's XNA Game Studio Express, which is a collection libraries and tools that facilitate game development. While development will take place on PCs, we will ultimately deploy our games on the XBox 360 console. At the end of the course we will present our results to the public. | |||||
| Lecture notes | Online XNA documentation. | |||||
| Prerequisites / Notice | The number of participants is limited. Prerequisites include: - good programming skills (Java, C++, C#, etc.) - CG experience: Students should have taken, at a minimum, Visual Computing. Higher level courses are recommended, such as Introduction to Computer Graphics, Surface Representations and Geometric Modeling, and Physically-based Simulation in Computer Graphics. | |||||
| 251-0576-00L | Digital Signal and Image Processing | W | 5 credits | 2V + 1U | G. Székely, S. Hirsch | |
| Abstract | The lecture provides an introduction to basic methods of digital image signal processing covering the following major topics: linear shift invariant systems and their characterization, the Fourier transform, signals in the spatial and frequency domain as well as sampling, quantization and interpolation. Theoretical and implementational issues about 1D and 2D FIR and IIR filters are also discussed. | |||||
| Objective | The goal of the lecture is to provide an introduction to basic knowledge and methods of signal processing, which is necessary to follow subsequent courses in visual computing (like computer graphics, computer vision or pattern recognition). While mostly concentrating on the processing of higher dimensional signals (2D, 3D), the course will be self-contained and discusses the underlying concepts also for the 1D (time-dependent) case. | |||||
| Content | The goal of the lecture is to provide an introduction to basic knowledge and methods of signal processing, which is necessary to follow subsequent courses in visual computing (like computer graphics, computer vision or pattern recognition). While mostly concentrating on the processing of higher dimensional signals (2D, 3D), the course will be self-contained and discusses the underlying concepts also for the 1D (time-dependent) case. Only basic concepts of real and complex analysis and probability theory will be assumed to be known. The course is given in English, but German can also be used for questions during the lectures, for the exercises and at the exams. Week 01: Discrete systems Week 02: Distributions, Linear, shift invariant systems (definition, characterization), convolution, harmonic waves Week 03: LSI description in spatial and frequency domains, the Fourier Transformation Week 04: Properties of the Fourier Transform, symmetries between the spatial and frequency domain, the Convolution Theorem and the Parseval Theorem, Windowing Week 05: Image Acquisition, Radiometric Calibration, Gamma Correction Week 06: Sampling, Shannon Theorem, Signal reconstruction Week 07: Interpolation, Quantization, Dithering Week 08: Histogram, intensity transformations, histogram manipulations (equalization and enforcement) Week 09: Finite Impulse Response (FIR) filters, their properties, basic filter types Week 10: Characterization of and design methods for FIR filters in the spatial and frequency domain Week 11: Infinite Impulse Response filters in one and two dimensions Week 12: Reconstruction of signals from projections, the Radon transformation | |||||
| 251-0581-00L | Computational Photography and Video | W | 6 credits | 2V + 1U + 1A | M. Pollefeys, G. Brostow | |
| Abstract | Computational photography combines plentiful computing, digital sensors, modern optics, and smart lights to escape the limitations of traditional cameras and enables novel imaging applications, such as unbounded dynamic range, variable focus, resolution, depth of field, hints about shape, reflectance, lighting, and new interactive forms of photos that are partly snapshots and partly videos. | |||||
| Objective | ||||||
| 263-9000-00L | Information Processing with Neural Networks | W | 4 credits | 2V + 1U | J. Bernasconi | |
| Abstract | Information processing with artificial neural networks (Basic principles and applications) | |||||
| Objective | The course gives an introduction to the different methods and techniques of information processing with artificial neural networks. Its aim is to provide the necessary background for an efficient use of these new information processing techniques. | |||||
| Content | Artificial neurons, different types of neural network paradigms (feedforward networks, Hopfield networks, winner-take-all networks), learning procedures (error backpropagation, stochastic learning, reinforcement learning, competitive learning), analysis and optimization of learning and generalization behavior, discussion and analysis of applications. | |||||
| Lecture notes | Course script (including a list of further references). | |||||
| Literature | Course script (with additional literature references) | |||||
| 402-0806-00L | Computational Vision | W | 6 credits | 2V + 1U | R. J. Douglas, D. Kiper, K. A. Martin | |
| Abstract | This course focuses on neural computations that underlie visual perception. We study how visual signals are processed in the retina, LGN and visual cortex. We study the morpholgy and functional architecture of cortical circuits responsible for pattern, motion, color, and three-dimensional vision. | |||||
| Objective | This course considers the operation of circuits in the process of neural computations. The evolution of neural systems will be considered to demonstrate how neural structures and mechanisms are optimised for energy capture, transduction, transmission and representation of information. Canonical brain circuits will be described as models for the analysis of sensory information. The concept of receptive fields will be introduced and their role in coding spatial and temporal information will be considered. The constraints of the bandwidth of neural channels and the mechanisms of normalization by neural circuits will be discussed. The visual system will form the basis of case studies in the computation of form, depth, and motion. The role of multiple channels and collective computations for object recognition will be considered. Coordinate transformations of space and time by cortical and subcortical mechanisms will be analysed. The means by which sensory and motor systems are integrated to allow for adaptive behaviour will be considered. | |||||
| Content | This course considers the operation of circuits in the process of neural computations. The evolution of neural systems will be considered to demonstrate how neural structures and mechanisms are optimised for energy capture, transduction, transmission and representation of information. Canonical brain circuits will be described as models for the analysis of sensory information. The concept of receptive fields will be introduced and their role in coding spatial and temporal information will be considered. The constraints of the bandwidth of neural channels and the mechanisms of normalization by neural circuits will be discussed. The visual system will form the basis of case studies in the computation of form, depth, and motion. The role of multiple channels and collective computations for object recognition will be considered. Coordinate transformations of space and time by cortical and subcortical mechanisms will be analysed. The means by which sensory and motor systems are integrated to allow for adaptive behaviour will be considered. | |||||
| Literature | Books: (recommended references, not required) 1. An Introduction to Natural Computation, D. Ballard (Bradford Books, MIT Press) 1997. 2. The Handbook of Brain Theorie and Neural Networks, M. Arbib (editor), (MIT Press) 1995. | |||||
| 252-4201-00L | Seminar Computational Geometry | W | 2 credits | 2S | B. Gärtner, E. Welzl, M. Hoffmann | |
| Abstract | This seminar is held once a year and complements the course Computational Geometry. Students of the seminar will present original research papers on computational geometry, some of them very recent. The seminar is a good preparation for a master, diploma, or semester thesis in the area of computational geometry. | |||||
| Objective | Each student is expected to read, understand, and elaborate on a selected research paper. To this end, (s)he should give a 45-min. presentation about the paper. The process includes * getting an overview of the related literature; * understanding and working out the background/motivation: why and where are the questions addressed relevant? * understanding the contents of the paper in all details; * selecting parts suitable for the presentation; * presenting the selected parts in such a way that an audience with some basic background in computational geometry can easily understand and appreciate it. | |||||
| Content | Computational Geometry is about design and analysis of efficient algorithms for geometric problems, typically in low dimensions (2,3,..). These are needed in many application domains, such as geographic information systems, computer graphics, or geometric modeling. | |||||
| Literature | Research papers as listed on the course webpage. | |||||
| Prerequisites / Notice | Prerequisites: participation (exam passed) in the course "Computational Geometry". A comparable course, for instance, attended at another university may also qualify; please contact the lecturers in such a case. Successful participation in the seminar requires the following: 1. a rehearsal talk, to be given in front of your supervisor at least one week prior to the plenary talk; 2. a satisfactory plenary talk; 3. attendance at all other talks. | |||||
| 251-0456-00L | Approximate Methods in Geometry Does not take place this semester. | W | 5 credits | 2V + 1U | E. Welzl | |
| Abstract | The course is concerned with approximate geometric methods for the analysis of large data sets represented by point clouds. Concrete topics are Low Distortion Embedding, Approximate Nearest Neighbor Search, Semi Definite Programming, Approximations and Nets, Approximate Smallest Enclosing Balls and Boxes, Directional Width, Support Vector Machines. | |||||
| Objective | ||||||
| Content | The course is concerned with approximate geometric methods for the analysis of large data sets represented by point clouds. Data is being collected in order to draw conclusions from it, i.e. to discover relations, make extrapolations into the future, etc. More often than not, data comes as a set or sequence of points describing objects, with each coordinate representing some quantification of some feature. On a computer such data is just a sequence of 0's and 1's; the need to analyze and "understand" it calls for means to support the process. One way is to visualize the data. For example, a data set representing a number of people by their respective heights and weights can be drawn as a point set in the plane, and this drawing may reveal some correlation that could be approximated by a linear function. For a wide range of applications (brain research, robotics, statistics, bioinformatics, character and speech recognition, computer graphics etc.) this approach is too simplistic, for various reasons. First of all, the size of the data may be huge (in the millions and billions, and sometimes so huge that we cannot even store it). And secondly, objects may have many features, giving rise to sets of dimension in the hundreds - and we know that simple visualization methods tend to fail starting in dimension 4. (Many features may in fact be redundant, but it is part of the endeavour to find out which ones.) Many of the arising problems appear to be too hard to be solved exactly in an efficient fashion. The course discusses several approximate methods for the analysis of large high-dimensional data sets that have been developed over the last years in response to the issues indicated above. While we have applications in mind for the questions we address, we emphasize theoretical aspects in the solutions (in particular, methods with guaranteed performance bounds). Methods we cover are random sampling, grid structures, core-sets, well separated pair decomposition, low distortion low-dimensional embeddings. Problems we address are shape and dimension analysis, nearest neighbor search, clustering etc. Examples for specific questions arising in these applications are the following: For some point in d-space, what is its closest neighbor in the point cloud? What is the closest pair in the point cloud? What is the "best" grouping of the points in the cloud into k groups? Which subset of the point set of size k provides the "best" description of the point cloud? What is the dimensionality of the point cloud and what does dimensionality mean here? Can the point cloud be embedded into a lower-dimensional space while preserving many of its characteristics? | |||||
| Prerequisites / Notice | By default the course will be given in German, but can be offered in English on demand. | |||||
Page
1
of
1
