Lecture notes : Algortihms used for Medical Image Processing

This course has two components. On the one hand, it is an introduction to digital image analysis presenting selected fundamental problems in medical image analysis, computer vision, photo/video editing, and graphics. We cover such basic concepts as image segmentation, registration, object recognition/matching, tracking, texture, etc. On the other hand, this is an applied course on standard computer science algorithms where students develop practical understanding of dynamic programming, graph based algorithms, computational geometry methods, etc. In fact, image analysis provides a stimulating environment for studying algorithms as their outputs can be intuitively visualized. Students with previous background in algorithms will be exposed to applications in image analysis, while students already familiar with problems in imaging will learn efficient methods based on standard CS algorithms. The course emphasizes the design, analysis, and implementation of algorithms in the context of 2D/3D medical images, photo and video data.

Lecture Notes( Download by Clicking the Topic Number)

  • Topic 2. Overview of different image modalities: photo images, video, and 2D-3D-4D medical data
  • Topic 3. Overview of basic image processing (point and local neighborhood processing): gamma correction, histogram equalization, window-center adjustment, linear filtering, image gradients.
  • Topic 4. Basic image segmentation in 2D (thresholding, region-growing, mean-shift, live-wire).
  • Topic 5. Deformable models (snakes): gradient descend, DP-snakes. Also distance transforms and generalized distance transforms.
  • Topic 6. Beyond snakes: implicit vs. explicit representation of boundaries, level-sets, geodesic active contours, geometric energy functionals.
  • Topic 8. Surface extraction in 3D, binary labeling, and graph cuts: volume segmentation, multi-view reconstruction, implicit vs. explicit boundary representation, binary submodular energies, geometric functionals, Markov Random Fields.
  • Topic 9. Multi-label image analysis problems: image restoration, stereo, texture synthesis, multi-object segmentation, pair-wise interaction potentials (convex vs. discontinuity preserving), energy minimization algorithms (simulated annealing, ICM, Ishikawa’s algorithm, multi-way graph cuts, a-expansions).


Related Posts Plugin for WordPress, Blogger...

Leave a Reply

Your email address will not be published. Required fields are marked *