1. Home
  2. NOVA School of Science and Technology
  3. Master in Biomedical Engineering
  4. Medical Image Processing

Medical Image Processing

Objectives

At the end of this course students should:
  • Know the specificities of a medical image, as a bi- or tri-dimensional signal
  • Be able to characterize an image as a digital signal, discretized in space and intensity
  • Know how to
    • characterize and evaluate the quality of an image
    • identify and remove common artefacts
    • apply digital filters to improve the quality of images
    • detect regions of interest
    • reconstruct images from projection, and its relation to the Radon transform
    • code and compress images
  • Understand how images are used diagnostics and to support medical decision making

General characterization

Code

12581

Credits

6.0

Responsible teacher

Ricardo Nuno Pereira Verga e Afonso Vigário

Hours

Weekly - 4

Total - 56

Teaching language

Português

Prerequisites

Students should have basic knowledge of signal processing. It would also be advisable to have similar knowledge level of the physical principles behind the generation of the most prominent medical imaging techniques.

Bibliography

  • Biomedical Image Analysis; Ragaraj M. Rangayyan; CRC Press, 2005

  • Digital Imaging Processing; Rafael C. Gonzalez and Richard E. Woods; Prentice-Hall, 2002

  • Digital Imaging Processing using Matlab; Rafael C. Gonzalez, Richard E. Woods and Steven L. Eddins;

    Prentice-Hall – New Jersey, 2004

  • Digital Signal Processing for Medical Imaging using Matlab;

  • Medical Image Analysis; Atam P. Dhawan; John Wiley & Sons, 2011

Teaching method

The theoretical lectures will be delivered by the lecturer, using the support materials that are deemed necessary for each topic. These will include the black board, slides, applets, and practical demonstrations demonstrations.
As practice training, the students will carry out a laboratory project, related to one particular problem of image analysis covered during the course. 

Evaluation method

Evaluation comprised two components, theory and practice:

  • Practical / "frequency": based on individual work (scripts distributed by the lecturer) and in groups of two students (a research project) –  frequency attained if the grade > 9.5. The grade is given after discussion with the docent;
  • Theoretical: 2 tests, carried out throughout the semestre. (to pass, this part of the evaluation needs to be > 9.5). Both tests have equal weight in the final grade.

 

The final grade will be calculated using the formula:

      (Final grade) = (tests)x0.7 + (frequency)x0.3

Subject matter

  1. The nature of biomedical images 

    1.1. data structure

    1.2. examples of the most common types of medical images and theur characteristics

  2. Image quality and information content

    2.1. characterization image quality

    2.2. spatial and intensity digitization

    2.3. optical density, dynamic range, contrast and resolution 

    2.4. histograms and entropy

    2.5. Fourier transform and spectral content

  3. Identification and removal of artefacts

  4. Image enhancement, through techniques based on

    4.1. averaging

    4.2. gray-scale transformations (Gamma, histograms...) 

    4.3. convolution and spatial filters

  5. Detection of regions of interest

  6. Image reconstruction from projections

  7. Image coding and data compression

  8. Pattern classification and diagnostic decision