Mathematical Imaging Working Seminar

Upcoming seminar

Feb. 29

Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Xianqi Li

Title: Iterative Sparse Maximum Likelihood - Based Algorithms
with Applications to SAR Imaging

Abstract: In this talk I will present a series of iterative sparse
maximum likelihood-based approaches (SMLA).

By using a particular form of Gaussian signal prior, iterative analytical expressions of
the signal and noise power estimates are obtained by iteratively
minimizing the stochastic maximum likelihood (SML) function with respect to only one scalar parameter
at a time, resulting in power-based SMLA approaches. However, these power-based sparse approaches do
not provide the phases of the unknown signals. To address this problem, a combined SMLA and Maximum
A Posteriori (MAP) approach (referred to as the SMLA-MAP approach) for estimating the unknown
complex-valued signals is proposed.

Finally, numerical examples of SAR imaging using Slicy data are
generated to compare the performances of the proposed and existing approaches.

Previous seminars


Spring 2011

Feb. 1, 8, 22

Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Professor Ling Pi

Title: Review of paper: Probabilistic inversion of a terrestrial ecosystem model:
Analysis of uncertainty in parameter estimation and model prediction (Author: Tao Xu et al.)

Jan. 18
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Professor Su-Shing Chen

Title: Systems Biology = Bioinformatics + Biomathematics? Process,
function and sequence

Abstract: I will describe this equation by a biological process- cell
cycle in dynamical systems, gene regulation networks, and image
processing. What we can solve using modewrn sciences- statistics,
mathematics and computer science.

Fall 2011

Nov. 30

Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: James Hungerford

Title: A Continuous Formulation of the Vertex Separator Problem (Continued)

Nov. 28
Time & Location: 5th period (11:45AM-12:35AM), in LIT233

Speaker: Xiaojing Ye

Title: Wiener Chaos Expansion and Numerical Solutions to Stochastic Differential Equations

Abstract: In this talk I will introduce the Wiener Chaos expansion method and
its application in solving randomly forced equations from a variety of
physics problems. The main advantage of the Wiener Chaos approach is
in separation of random and deterministic effects of the random force
in a rigorous and effective manner. We show that this method is more
efficient and accurate than those based on the Monte Carlo
simulations.

Nov. 16
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: James Hungerford

Title: A Continuous Formulation of the Vertex Separator Problem

Abstract: Given an undirected graph G, the vertex separator problem
is to find the smallest number of nodes whose removal disconnects
the graph into two subsets which satisfy specified size constraints.
We show the VSP can be formulated as quadratic program for which
local optimality can be checked quickly.

Nov. 9
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Prof. Min Zhou

Title: The Research of Theory and Technology in Subdivision Surface Modeling (Continued)

Nov. 2
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Prof. Min Zhou

Title: The Research of Theory and Technology in Subdivision Surface Modeling

Topics:
    1. Geometric modeling technology
        ·Polygonal meshes modeling technology
        ·Implicit expression equation surface modeling technology
        ·Parametric surface modeling technology
        ·Subdivision surface modeling technology
    2. Analysis of the research actualities on tubdivision method
    3. The recent research work of our group
        ·Fundamental theory of subdivision inverse problem
        ·High-performance subdivision algorithm based on subdivision inverse problem
        ·The method of subdivision inverse problem

Oct. 26
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Iulia Posirca

Title: Variational models for multiplicative noise removal

Abstract: It is well known that multiplicative noises are commonly found in real world image processing applications,
such as SAR images, laser images and medical ultrasonic images. Unlike additive noises, the multiplicative ones
are more difficult to be removed from the corrupted images. One reason is related to their multiplicative nature.
In plus, the distribution of such noises is not Gaussian. In this talk, I will review some variational models for removing
the multiplicative noise.

Oct. 12
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Meng Liu

Title: A Fast Algorithm for MR Image Reconstruction in Partially Parallel Imaging (Continued)

Oct. 5
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Meng Liu

Title: A Fast Algorithm for MR Image Reconstruction in Partially Parallel Imaging

Abstract: A fast numerical algorithm will be presented for solving total variation and L1 based image reconstruction with application in partially parallel MR imaging (PPI). The algorithm uses variable splitting method to reduce computational cost. Moreover, the Barzilai-Borwein step size selection method is adopted in the algorithm for much faster convergence.

Sep. 28
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Fuhua Chen

Title: Graph-cut based image segmentation (Continued)

Sep. 21
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Fuhua Chen

Title: Graph-cut based image segmentation (Continued)

Sep. 14
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Fuhua Chen

Title: Graph-cut based image segmentation (Continued)

Sep. 7
Time & Location: 4th period (10:40AM-11:30AM), in LIT305

Speaker: Fuhua Chen

Title: Graph-cut based image segmentation

Abstract: Graph-cut method has been widely used in computer
vision, including image segmentation. In this talk, we will
introduce how graph-cut can be used to image segmentation, the
extension from discrete graph-cut (traditional one) to continuous
graph-cut and the extension from two-phase segmentation to
multiphase segmentation.