COMP 775: Introduction to Medical Image Analysis

Fall 2008, M W 2:00pm-3:15pm, SN 011

Instructor: Marc Niethammer
Email: (mn -at- cs.unc.edu)
Phone: 919-843-7449
Office:219 Sitterson Hall

Quick links: schedule, useful resources

Course description [PDF]

Overview

This course will provide an introduction to medical image analysis. Starting with an overview of imaging modalities, we will explore simple methods for image preprocessing and feature extraction, methods for image segmentation, for image analysis, as well as for image registration. The course will only require minimal prerequisites, mainly some knowledge in linear algebra and calculus.

Prerequisites: Some knowledge of linear algebra and calculus.

Resources: A textbook is not explicitly required. However, I will follow loosely material in the book 'Image Processing, Analysis, and Machine Vision' by Sonka, Hlavac, Boyle, Third Edition.

Lecture materials will be posted on this web page as well as on blackboard

Grading: Grades will be based on one take-home mid-term exam, a few homework problems, and a small course project. The split-up is as follows:
  • Mid-term: 30%
  • Homework: 30%
  • Project: 40%

Schedule (in flux)

Date Topic Notes
Mon Aug. 18 No class: Computer Science Research Fair  
Wed Aug. 20 History of Medical Imaging. Overview of modalities: microscopy, ultrasound, x-ray, CT, PET, MRI Overview of imaging modalities [PDF]
Mon Aug. 25 Simple feature extraction and preprocessing: smoothing, edges and corners, Hough transform. Smoothing, edges, corners [PDF]
Wed Aug. 27 Simple feature extraction and preprocessing ... continued.  
Mon Sep. 1 No class: Labor day  
Wed Sep. 3 Binary images: Morphological processing, skeletonization, ... Morphology, etc. [PDF]
Mon Sep. 8 Segmentation: Thresholding and edge-based methods. Simple segmentation/thresholding methods [PDF]
Wed Sep. 10 Segmentation: Merging, splitting, watershed, and matching. MICCAI: Need to reschedule.
Mon Sep. 15 Bayesian estimation, ML, MAP, ... Optimality and Bayesian estimation [PDF]
Wed Sep. 17 Bayesian estimation, ML, MAP, ... continued.  
Mon Sep. 22 Estimation of probability densities.  
Wed Sep. 24 Segmentation through classification and the Bayesian rationale for energy functionals. Energy-based segmentation [PDF]
Mon Sep. 29 Contour-based shape descriptions.  
Wed Oct. 1 Introduction to variational calculus. Calculus of variations [PDF]
Mon Oct. 6 Introduction to variational calculus ... continued.  
Wed Oct. 8 Active Contours and Snakes. Some variational segmentation approaches [PDF]
Mon Oct. 13 Dicussion of take-home midterm.  
Wed Oct. 15 Graph cuts. Graph cuts (contains movies) [PPT]
Mon Oct. 20 Orthogonal Function Decomposition. Representations [PDF]
Wed Oct. 22 Eigenanalysis, singular values, principal component analysis.  
Mon Oct. 27 Point distribution model, active shape and active appearance models, medial models. Active shape and active appearance models [PDF]
Wed Oct. 29 Point distribution model, active shape and active appearance models, medial models ... continued.  
Mon Nov. 3 Analysis: Simple shape descriptors.  
Wed Nov. 5 Classification through clustering. Clustering [PDF]
Mon Nov. 10 Classification through clustering ... continued.  
Wed Nov. 12 Basics on validation: overlap measures, experimental designs, ROC curves.  
Mon Nov. 17 Statistical testing.  
Wed Nov. 19 Registration: Transformation groups and similarity metrics. Registration [PDF]
Mon. Nov 24 Elastic registration: Optical flow.  
Wed. Nov 26 Elastic registration: Thin plate splines; outlook to fluid flow approaches.  
Mon Dec. 1 Applications for image registration: atlas construction, segmentation through registration, ...  
Wed Dec. 3 Wrap-up.  

Useful Resources