| Talk Date: | 20090121 |
| Speaker Name: | Alexander Petukhov |
| E-mail: | petukhov@math.uga.edu |
| webpage: | |
| Talk Title: | Underdetermined Linear Systems and Information Theory problems |
| Talk Abstract: | We discuss how underdetermined linear systems are connected with problems of data representation and transmission. The algorithms for finding sparse solutions and their sensitivity to data precision will be discussed. The considered new algorithm combines features of linear programming and orthogonal greedy algorithm and (to our knowledge) outperforms other algorithms. |
| Talk Date: | 20090128 |
| Speaker Name: | Ming-Jun Lai |
| E-mail: | mjlai@math.uga.edu |
| webpage: | www.math.uga.edu/~mjlai |
| Talk Title: | Bivariate Splines for Ozone Concentration Forecasting |
| Talk Abstract: | Abstract: This is based on a joint work with Serge Guillas and Bree Ettinger. We consider the functional linear regression models where the explanatory variable is a random surface and the response is a real random variable with bounded noise. Bivariate splines over triangulations represent the random surfaces. We use this representation to construct least squares estimators of the regression function and autoregressive estimators based on principal component analysis. We carry out an application of these two functional linear models to ozone concentration forecasting over the U.S. to illustrates the predictive skills of these two methods. |
| Talk Date: | 20090204 |
| Speaker Name: | Bree Ettinger |
| E-mail: | bree@math.uga.edu |
| webpage: | |
| Talk Title: | Bivariate Splines for Ozone Concentration Forecasting (II) |
| Talk Abstract: | Abstract: This is based on a joint work with Serge Guillas and Ming-Jun Lai. We consider the functional linear regression models where the explanatory variable is a random surface and the response is a real random variable with bounded noise. Bivariate splines over triangulations represent the random surfaces. We use this representation to construct autoregressive estimators based on principal component analysis. We carry out an application of the functional linear models to ozone concentration forecasting over the U.S. to illustrates the predictive skills of the method. |
| Talk Date: | 20090211 |
| Speaker Name: | Andrew Sornborger |
| E-mail: | ats@math.uga.edu |
| webpage: | |
| Talk Title: | An Overview of the Singular Value Decomposition and Its Use |
| Talk Abstract: | The singular value decomposition is one of the most important decompositions used in applied mathematics. I will derive the decomposition and then discuss typical ways in which it is used with an example from biological imaging. |
| Talk Date: | 20090225 |
| Speaker Name: | Leopold Messi |
| E-mail: | |
| webpage: | |
| Talk Title: | How to make a poster |
| Talk Abstract: | Many conferences offer a poster session. I will explain how to use TeXnology to make a poster of any size. The poster can be printed at FedExKinko with a reasonable price. Recently I made it with mail tube for $8. |
| Talk Date: | 20090318 |
| Speaker Name: | George Yin |
| E-mail: | gyin@math.wayne.edu |
| webpage: | |
| Talk Title: | Switching Diffusion Processes: Some Recent Progress |
| Talk Abstract: | In this talk, we report some of our recent work on switching diffusion processes in which continuous dynamics and discrete events coexist. First, motivational examples arising from manufacturing, control and optimization of stochastic systems, insurance and finance, singular perturbed Markovian systems will be mentioned. After recalling the notion of recurrence and regularity, necessary and sufficient conditions for recurrence are provided; ergodicity will be examined; stability will be studied. |
| Talk Date: | 20090325 |
| Speaker Name: | Alexander Petukhov |
| E-mail: | petukhov@math.uga.edu |
| webpage: | |
| Talk Title: | Sparse representations and digital film restoration. |
| Talk Abstract: | Most of methods of signal processing explicitly or implicitly assume separation of useful signal or artifacts have some sparse representation. Digital film restoration provides us with very wide spectrum of signal processing problems obeying the general principles of sparse representation. At the same time, they are very different and require different approaches. Among those problems we consider 1) The problem of dust/dirt removal. It has some feature of Error Correcting Code. 2) Film grain removal. It is similar to classical denoising problems. 3) Flicker removal. The film archives community requires to leave "healthy" information unchanged. Compare to classic requirement to maximize signal-to-noise ratio. For example, such classical methods as the Wiener filtration do not satisfy that requirement. Besides, the dimension of those problems is very high. So the sparsity is not only the way to separate data and artifact but sometimes is the only way to provide computationally efficient algorithm. All those problems are heavily dependable on the motion estimation. Which is critically important for any video processing. We also are going to address the issue of motion estimation. We will demonstrate some samples of restoration of unique materials from Library of Congress, Russian National Film Archive. Among those samples we will show a paper print (probably produced by T.Edison), Leo Tolstoy funeral, A.Schweizer plays Bach. |
| Talk Date: | 20090401 |
| Speaker Name: | James Nagy |
| E-mail: | nagy@mathcs.emory.edu |
| webpage: | www.mathcs.emory.edu/~nagy |
| Talk Title: | Large Scale Inverse Problems in Imaging |
| Talk Abstract: | Digital images are used to analyze objects in a variety of applications, such as star clusters in astronomy, molecules in biology, and tumors in medicine. Often postprocessing must be done to the collected data; this may involve reconstructing, restoring or enhancing the image. Mathematically these processes are modeled as inverse problems. Inverse problems usually cannot be solved analytically, and thus computational approaches must be used. In imaging, these computational problems are very large, requiring development of efficient solvers. Another difficulty is that the problems are very sensitive to errors, such as noise, in the data. This difficulty is usually handled by a technique called regularization. In this talk we describe some inverse problems that arise in imaging applications, approaches to regularize them, and our recent contributions to the development of iterative solvers. |
| Talk Date: | 20080408 |
| Speaker Name: | Sung Ha Kang |
| E-mail: | kang@math.gatech.edu |
| webpage: | http://www.math.gatech.edu/~kang/ |
| Talk Title: | Variational Models for Image Colorization via Chromaticity and Brightness Decomposition |
| Talk Abstract: | Colorization refers to an image processing task which recovers color of gray scale images when only small regions with color are given. We propose a couple of variational models using chromaticity color component to colorize black and white images. We first consider Total Variation minimizing (TV)colorization which is an extension from TV inpainting to color using chromaticity model. Secondly, we further modify our model to weighted harmonic maps for colorization. This model adds edge information from the brightness data, while it reconstructs smooth color values for each homogeneous region. We introduce penalized versions of the variational models, we analyze their convergence properties, and we present numerical results including extension to texture colorization. |
| Talk Date: | 20090422 |
| Speaker Name: | Weihong Guo |
| E-mail: | wguo@as.ua.edu |
| webpage: | http://www.bama.ua.edu/~wguo2/ |
| Talk Title: | Use patch information to remove noise and artifacts from medical images |
| Talk Abstract: | Medical images are usually polluted by noise and artifacts. Fast imaging technologies which involve under-sampling lead to images with more spatially inhomogeneous noise and artifacts. This talk is about our recent work on using local patch information to efficiently and automatically improve image qualities. It is based on a joint work with Feng Huang. |
| Talk Date: | 20090429 |
| Speaker Name: | Caner Kazanci |
| E-mail: | cmeile@uga.edu |
| webpage: | www.math.uga.edu/~caner |
| Talk Title: | Stochastic simulation algorithms for modeling biological and ecological systems |
| Talk Abstract: | Stochastic simulation methods are essential for modeling biological and ecological systems. Unfortunately few software offers powerful stochastic solvers. We will start with a few examples demonstrating cases where a stochastic simulation is needed. These cases are characterized by a significant differences between the ODE and SDE solutions. Then, we will go over Basic Stochastic Algorithm, followed by Gillespie's stochastic algorithm, including examples for each method. We will than derive the master equation, which describes the probability density function of the state of the system. Master equation converges to a Fokker-Planck type PDE, which is associated with an SDE. Called "Chemical Langevin Equation", this will be the third (and fastest) stochastic method we will cover. We will also find out why the ODE and SDE solutions differ significantly in some cases. I will provide simple pseudo-codes and outputs for each stochastic method. |
| Talk Date: | 43723634551 |
| Speaker Name: | vaksdpowo 462526 |
| E-mail: | imsaet@cgftje.com |
| webpage: | http://bfymikzwqhof.com/ |
| Talk Title: | bBlQuuNbbUkC |
| Talk Abstract: | TUBwkj mndnoulpdbnt, [url=http://sopmwlmiovfo.com/]sopmwlmiovfo[/url], [link=http://umccxdpqebgt.com/]umccxdpqebgt[/link], http://kfsmcidjbxho.com/ |