ANNOUNCEMENTS |
Course will be offered when demand is large enough. If you are interested in this course, please let me know.
SUMMARY |
Gain in-depth understanding, knowledge, and the ability to apply, cutting-edge methods to process and quantitatively analyze images. This class covers special and advanced topics of modern image analysis including wavelets, adaptive filters, active contours, and fractals. An important aspect is the design of unsupervised image analysis chains.
Prerequisites: ENGG 4620/6620 or instructor's consent.
Class schedule Fall 2017:
TBD
CONTENTS |
INTRODUCTION |
In ENGG4620, we examined how medical images are generated using CT, MRI,
ultrasound, and other techniques. Since these modalities rely on computerized
data processing, fundamental image processing steps were introduced.
What happens with those images? All too often, the radiologist hits
the PRINT button, exposes a film and pops it into his light box for viewing.
But the computer can do more!
The computer can improve the images by eliminating noise, amplifying
contrast or accentuating features to be analyzed (enhancement).
Medical image analysis, however, is a complex task, calling for
complex algorithms. With ever increasing computer power, very sophisticated
algorithms have been introduced in recent years. In this class, we will cover some of
those advanced algorithms.
The goal of this class is to give you knowledge and skills as follows:
SYLLABUS
Please note that this table is tentative
and will be updated as we go.
BOOKS
Recommended textbook for this class is:
M.A. Haidekker: Advanced Biomedical Image Analysis. Wiley 2011, ISBN 9780470624586.
You may also opt for any of these alternative textbooks:
Computer Imaging: Digital Image Analysis and Processing by Scott E Umbaugh
(Taylor & Francis)
Medical Image Analysis by Atam P. Dhawan (IEEE Press)
COMPUTER LAB
Alternatively, you may obtain the DVD that accompanies the textbook.
This is a live DVD that allows you to run the software without modifying your
hard drive.
GRADING
The grade will be based on the exam-style project, the homeworks,
and the student lecture.
You will receive score points based on the fill-the-bucket principle,
i.e. for each homework assignment and for each test, you accrue score points.
Your final grade will be determined from the score you achieved relative
to the maximum score achievable. Typically, you receive a maximum of 15
points per homework, 100 points for the exam, and 100 points for
the presentation, resulting in a maximum score of around 300 points.
We use a fixed grading system. There will be no adjustment
based on the overall class performance.
The following table shows the percentage of your score you need to reach
for a specific grade:
CLASS SCHEDULE AND OFFICE HOURS
Class schedule TBD:
Office hours are by appointment.
The best way to contact me is email.
HOMEWORKS
STUDENT LECTURES AND PROJECTS
The following student lectures have been assigned:
LINKS
The computer can separate features from background (segmentation).
The computer can take objective measurements of features, thus
aiding the radiologist in his diagnosts (quantification)
The computer can superimpose or merge images from different modalities
(registration)
The computer can reduce the data size for improved storage (compression)
And the computer can display image data in various ways (visualization).
Lecture Block Topic
Block 1 Survey of fundamental image processing steps:
Enhancement, segmentation, quantification.
Block 2 The design of an image processing chain.
Supervised, user-aided, and unsupervised operations.
Blocks 3 and 4 The Fourier transform, discrete cosine transform, and Hartley transform. Frequency-domain filtering. Wiener filtering.
Blocks 5 and 6 The wavelet transform and wavelet-based filtering
Block 7 Adaptive filters in the spatial and frequency domain.
Block 8 Active contours and deformable models
Block 9 Segmentation and object recognition with the Hough transform
Block 10 Clustering techniques and multispectral image segmentation
Block 11 Texture classification and quantification
Block 12 Shape extraction, shape quantification and classification
Block 13 Fractal approaches to image quantification
Block 14 2D and 3D image visualization
Block 15 Trends and current developments in image processing
Grade Minimum percentage
Grade Minimum percentage
Grade Minimum percentage
A+ 96%
A 93%
A- 90%
B+ 85%
B 80%
B- 75%
C+ 70%
C 65%
C- 60%
D+ 50%
D 45%
No.
Topic
Due date
Name
Lecture topic
Lecture date