CMPT 888: Medical Image Analysis (appears under: Special Topics
in Computer Graphics, HCI, Computer Vision, and Visualization)
Next offering: TBD.
Note to non-SFU graduate
- General objectives of the course: Through round-table discussions
and presentations of classical and state-of-the-art papers on medical image
analysis, we will review and critique how different computational,
engineering, and mathematical techniques are adapted and applied to the
problem of extracting clinically useful information from medical images.
- In this course you will:
- learn about medical imaging modalities such as MRI, CT, and ultrasound,
and recent advances such as functional and molecular imaging, diffusion
tensor imaging, and time-varying medical images.
- appreciate the need for developing automated, accurate, robust, and fast,
techniques for extracting information from medical imaging for a variety of
- learn how the field of medical image analysis involves understanding,
adapting and applying a variety of computational, mathematical and
engineering techniques which include: signal/image processing, optimization,
AI, machine learning, graph theory, mathematical modeling, differential
equations, multi-variate statistics, geometrical modeling,
matrix/tensor algebra, etc.
- learn about classical and contemporary techniques for medical image
segmentation, medical image registration, and shape analysis, which
constitute three important sub-areas of medical image analysis.
- learn what medical image segmentation is and why is it important for
medicine, understand some of the most important established and researched
techniques for segmentation, and appreciate some of the outstanding
- understand what medical image registration is, its applications in
health care, the most established frameworks for registration, and
what some of the remaining challenges are.
- learn about different ways of representing and analyzing anatomical
structures of deformable shapes and how can this translate into improved
- Project: The course involves working on a medical image analysis
project. A wide range of projects related to the above topics is possible.
The project topic can be either initiated by the student or recommended by
the instructor. The project may be very clinical application oriented
or may deal with a theoretical aspect of computational/mathematical
techniques for medical image analysis, or a combination of both (see some
projects at our Medical Image
- Advantages: there are many other advantages to studying topics at
the interface between computing and medicine, including: wider career
opportunities, graduate studies and funding, working with state-of-the-art
software and hardware, collaborating with doctors and accessing valuable
medical data... more are mentioned here.
vessel segmentation from 3D MR brain angiography (head seen from above, nose
in lower right corner)
segmenting the spinal cord from a mid-sagittal MRI
segmentation of the rotator cuff muscle in the shoulder
Registration of anatomical CT image with data from a functional SPECT scan
Manifold Learning and graph representation of medial (skeleton) based
Decomposing shape variability of a brain structure (corpus callosum)
into localized and intuitive deformation