Teaching     Home

CSE5392 (Spring 2007)
Hot Topics of Computer Science Applications to Medical Imaging

Heng Huang


[ Administrative Basics | Course Description | Assignments | Outline of Lectures ]

Administrative Basics

Lecture

Nedderman Hall 315 | Tuesdays and Thursdays 12:30-1:50 PM
Instructor

Heng Huang | Nedderman Hall 308 | Office hours: Tuesdays and Thursdays 2:00-3:00 PM
Textbook

Required: none

Recommended:
Insight into Images: Principles and Practice for Segmentation, Registration, and Image Analysis. Edited by Terry S. Yoo. August 2004; ISBN 1-56881-217-5 .

Good books for reading:
Foundations of Medical Imaging, Z. H. Cho, Joie P. Jones, Manbir Singh. 1993; ISBN: 978-0-471-54573-6.
Digital Image Processing. Rafael C. Gonzalez, Richard E. Woods. 2002; ISBN: 0201180758 .
Work

Two homework sets. (20%)
Class presentations. (20%)
Writing final reports (that will be posted in this website) or
using alternate programming projects (a short presentation is required in class). (40%)
Participation. (20%)
Request

Basic math and programming background, for undergraduate student, CSE 2320, CSE 4303 or CSE 4313, are requested

Course Description

The aim of the course is to show how computer science can be used to model and analyze medical data in order to provide a prognosis and develop cures. Medical image computing is, by nature, an interdisciplinary field involving not only medicine but also computer science, mathematics, biology, psychology, statistics and other fields. The "glue" to all is computer science which can help detect patterns and make sense out of disparate types of information.

The course is application-driven and includes topics in medical image analysis, including image segmentation, registration, statistical modeling and applications of computational science in enabling treatment. It will also include selected topics relating to medical image acquisition, especially where they relate to analysis. The course will provide the participants with a thorough background in current research in these areas, as well as to promote greater awareness and interaction between multiple research groups within the university.

Assignments

Homework 1 (Please send me your solutions of HW1 before Mar. 28)
Homework 2 (Please send me your solutions of HW2 before Apr. 26 or in class)

Final Programming Project -- segmentation medical images using "snakes"

Project Requirements:
MATLAB - Programming language
Writing code by yourself

Project Description:
Snakes are computer-generated curves that move within images to find object boundaries. They are often used in medical image analysis to detect and locate objects, and to describe their shape. Rough shape and starting position of the snake are specified by the user. This is done by clicking a few points on the image which become the vertices of the initial snake. The snake is defined as an energy minimising contour where the energy function is a combination of internal and external forces. An iterative procedure causes the snake to shrink, reduce its total curvature and move towards interesting objects in image. In our class, we have studied the regular snake method. Please use MATLAB to implement the regular snake method for medical image segmentation. The testing medical images will be sent to students by email.

Extra Credit:
We have discussed in class about using dynamic programming method to fast minimize the energy function. Please use this method to improve your snake model. The image should be considered as eight-connection image.

Due on May 1st.

Final Report Project

Please select one of the following topics as your final report:
1) medical image segmentation methods;
2) medical image registration methods;
3) medical image analysis in cancer detection;
4) medical image analysis in computer assisted surgery.
You should study and summarize several medical image analysis methods in your topic and talk about their advantages and disadvantages in different image modalities. No more than 10 pages.

Due on May 10th.


Outline of Lectures

Week 1.

Tue Jan 16: Introduction to Medical Image Processing, Course Objectives
Thu Jan 18: Basic Image Processing, Linear Operators

Week 2.

Tue Jan 23: Discrete Fourier, Discrete Cosine Transforms.
Reading materials: [1], [2] (including spectral analysis and filtering sections)
Thu Jan 25: Physics of Medical Imaging 1

Week 3.

Tue Jan 30: Physics of Medical Imaging 2
Reading materials: The Basics of MRI
Thu Feb 01: Physics of Medical Imaging 3

Week 4.

Tue Feb 06: Statistics of Pattern Recognition 1

Week 5.

Tue Feb 13: Statistics of Pattern Recognition 2
Thu Feb 15: PDE-based Nonlinear Image Filtering

Week 6.

Tue Feb 20: Segmentation for Medical Images 1 (edge detection, region growing, etc.)
Thu Feb 22: Segmentation for Medical Images 2 (snakes)

Week 7.

Tue Feb 27: Segmentation for Medical Images 3 (level set)
Thu Mar 01: Segmentation for Medical Images 4 (watershed) & Mathematical Morphology

Week 8.

Tue Mar 06: Medical Images Registration 1 (Intensity-Based Registration)
Thu Mar 08: Medical Images Registration 2 (Feature-Based Registration) + homework discussion

Week 9.

Spring Break

Week 10.

Tue Mar 20: Medical Images Registration 3 (Robust Estimation)
Thu Mar 22: Medical Images Registration 4 (Deformable Registration)

Week 11.

Tue Mar 27: Visualization 1 (Scalar Visualization)
Reading materials:
Marching Cubes: A High Resolution 3D Surface Construction Algorithm
The Asymptotic Decider: Removing the Ambiguity in Marching Cubes
Thu Mar 29: Visualization 2 (Volume Visualization)
Reading materials:
Back-to-Front Display of Voxel-Based Objects
H. Tuy and L. Tuy. Direct 2D Display of 3D Objects. IEEE Computer Graphics and Applications, 4(10):29--33, 1984 (please look at this paper at library).

Week 12.

Tue Apr 03: Class Presentations (An)
Improved Watershed Transform for Medical Image Segmentation Using Prior Information Slides
Solution for Homework 1
Thu Apr 05: No Class (I have a seminar talk at UTSW about biomedical image computing)
I will have an extra office hour on Apr. 6, 2pm - 4pm for homework and project.

Week 13.

Tue Apr 10: Class Presentations (Rebecca Atchley, Beatrice)
A Geometric Snake Model for Segmentation of Medical Imagery Slides
Unsupervised border detection in dermoscopy images Slides
Thu Apr 12: Class Presentations (Ankit, Nayan Asanani)
A Support Vector Machine Approach for Detection of Microcalcifications
Delineating Fluid-Filled Region Boundaries in Optical Coherence Tomography Images of the Retina Slides

Week 14.

Tue Apr 17: Class Presentations (Ignatius Gomes)
Imaging the Cardiovascular Pulse Slides
Lecture: Active Shape Model for Medicl Image Segmentation
Thu Apr 19: Shape Modeling and Analysis 1

Week 15.

Tue Apr 24: Shape Modeling and Analysis 2
Thu Apr 26: Class Presentations (Ramya, Nirupama)
Gray Scale Registration of Mammograms Using a Model of Image Acquisition
Diffuse optical imaging of brain activation: approaches to optimizing image sensitivity, resolution, and accuracy Slides

Week 16.

Tue May 01: Class Presentations (Sennen)
Spectral filter optimisation for the recovery of parameters which describe human skin Slides
Lecture: Biomedical Image Applications 1
Thu May 03: Biomedical Image Applications 2


Paper List for Presentation

Presentation:

Every student selects one paper from the following list and presents the paper (25 min ~ 40 min) in class. Grade: slides preparation 50%, oral presentation 25%, and answer questions 25%. The presentation on paper in IEEE Transactions on Medical Imaging will get plus.

Segmentation:

A Geometric Snake Model for Segmentation of Medical Imagery. (Rebecca Atchley)
An Adaptive Level Set Method for Medical Image Segmentation.
CLASSIC: Consistent Longitudinal Alignment and Segmentation for Serial Image Computing.
Robust Active Appearance Model Matching.
Simultaneous Segmentation and Registration of Contrast-Enhanced Breast MRI.
Multiscale Vessel Enhancing Diffusion in CT Angiography Noise Filtering.
Improved Watershed Transform for Medical Image Segmentation Using Prior Information. (Vo, An)

Registration:

Image Registration Based on Thin-Plate Splines and Local Estimates of Anisotropic Landmark Localization Uncertainties.
A method for the registration of 3-D shapes.
A Novel Parametric Method for Non-rigid Image Registration.
Multimodality Image Registration Using an Extensible Information Metric and High Dimensional Histogramming.
Gray Scale Registration of Mammograms Using a Model of Image Acquisition. (Ramya Ananthanarayanan)
A View-Based Approach to Registration: Theory and Application to Vascular Image Registration.

Others:

A Support Vector Machine Approach for Detection of Microcalcifications. (Ankit Master)
Imaging the Cardiovascular Pulse. (Ignatius Gomes)
Top-Down and Bottom-Up Strategies in Lesion Detection of Background Diabetic Retinopathy. (Rashida)
Scale Selection for Anisotropic Scale-Space: Application to Volumetric Tumor Characterization.
Spectral filter optimisation for the recovery of parameters which describe human skin. (Sennen Pinto)
An inverse method for the recovery of tissue parameters from colour images.
Delineating Fluid-Filled Region Boundaries in Optical Coherence Tomography Images of the Retina. (Nayan Asanani)
Unsupervised border detection in dermoscopy images. (Beatrice)
Diffuse optical imaging of brain activation: approaches to optimizing image sensitivity, resolution, and accuracy. (Nirupama)