Practical use of Motion Magnification

Micro-motions can reveal a lot of information about the underlying system. For example, with every heart pulse, your face changes slightly. This subtle change is not visible to the human eye. However, with the right image processing, these invisible motions can be magnified from an ordinary video. Other examples include the magnification of micro-vibrations in buildings or machines to avoid fatigue. This magnification of micro-motion from a video is called motion magnification. In this semester's project, you will develop and evaluate a so-called motion magnification algorithm. If time allows, you will investigate possible applications such as pulse estimation or seismic monitoring. 

  • Tasks

    1. Study the principles of motion magnification
    2. List different algorithms
    3. Implement an algorithm (primary goal)
    4. Qualify algorithm according to different metrics, such as image quality, video encoding, video length, shutter speed, etc. (secondary goal)
    5. Investigate potential applications (secondary goal)

  • Further Information

    • 20% theory, 80% implementation
    • Term paper, 1-2 students
    • Prior knowledge recommended in
      • Image and/or signal processing
      • C# or Python
    • Student(s) should be curious and motivated to learn a new technologies

  • References

    C. Liu, A. Torralba, W.T. Freeman, F. Durand and E.H. Adelson. Motion Magnification. Accepted by Siggraph, 2005. 

    Wadhwa, Neal, et al. "Riesz pyramids for fast phase-based video magnification." 2014 IEEE International Conference on Computational Photography (ICCP).

Have we aroused your interest?

I am interested in the student research project Practical use of Motion Magnification and would like to learn more. learn more.