| Recommended semester: 4th -7th semester |
| Scope and form: Project work in groups of 2 persons |
| Evaluation: Approval of coursework/reports
|
| Examination: Pass/fail |
| Previous course: 04401 |
| Prerequisites: Knowledge of introductory image analysis and numerical methods for optimization. |
| Participant limitation: Max. 20 |
| Aim: To give the students knowledge of some problems from image analysis, setting up mathematical (statistical) models, and the use of optimization algorithms to get robust estimates of the parameters in the models. |
Contents: After an introduction on the first day, the students work in groups of 2-4 persons. Each group has its own problem (which the members help to choose and define). The problems are taken from current research at IMM, e.g.
1. Feature matching: Find corresponding 'features' in two images and find the best match, taking into account the camera optics. Occurs e.g. in stereo vision.
2. Elektrophooresis Gel Matching (used to analyse proteins): Find an optimal distortion of the image of a gel to give a best match with another. This problem has similarities with 1.
3. Deformable models: Use prior assumptions and observations to find an optimal segmentation. Used to find the fat and lean parts in images of meat. |
Contact: Jens Michael Carstensen, building 321, (+45) 4525 3412, jmc@imm.dtu.dk
Hans Bruun Nielsen, building 305, (+45) 4525 3077, hbn@imm.dtu.dk |
| Department: 002 Informatics and Mathematical Modelling |
| Course URL: http://www.imm.dtu.dk/courses/02507 |
| Updated: 04-05-2001 |
