Web-Based CT Image Deblurring
Ming Jiang and Ge
Wang
Departments of Biomedical Engineering and
Radiology
University of Iowa
jiangm@ct.radiology.uiowa.edu
Introduction
Digital image restoration is the
established methodology to achieve super-resolution retrospectively.
Recent theoretical progress suggests that the expectation maximization
approach can be applied for optimal image deblurring. In our earlier
work, we experimentally validated the 3D Gaussian point spread function
(PSF) of the CT scanner. The spiral CT image deblurring algorithm was
formulated with the knowledge of the PSF. A typical cochlear
cross-section was used for evaluation, demonstrating a resolution gain
up to 30-40%. In practice, it is difficult or even impossible to
determine the PSF for every image to be deblurred. Blind deblurring is
a technique for image deblurring without the exact knowledge of the
associated PSF. We have recently achieved promising results of blind
deblurring.
Problem Statement
We want to further evaluate and refine our
blind deblurring technique and implement this unique post-processing
capability on the web so that all colleagues over the world can upload
their CT images and download substantially improved images.
Suggested Methods
Our image deblurring codes were developed
in C and MatLab, respectively. C/C++ is the most popular programming
language in the engineering field. Also, many professionals rely on
MATLAB to accelerate their research. The MATLAB environment encourages
creativity and enables you to quickly test alternatives. It is
desirable for you to know C/C++ and/or MatLab so that you can make use
of the available codes. For development of a web-based CT image
deblurring resource, you may also consider use of Java or other tools.
Expected Results
You are expected to understand the
principles of the CT image deblurring procedure, produce a set of
experimental data demonstrating the image deblurring effects, and most
importantly develop a web-based interface so that the deblurring
software is made publicly accessible.
CT Cross-Section of the Inner Ear Before (Left) and After (Right) Deblurring