Deformable Models in Medical Image Segmentation

Becker, M. and Magnenat-Thalmann, N.

Abstract: Today medical imaging techniques are very common and frequently used. To assist doctors and to perform automated analysis, the images have to be segmented by organ or region of interest. This segmentation process is a complex task. Relying on the image acquisition, segmentation approaches have to be robust and flexible enough to withstand low contrast, artifacts, small field of view and other phenomena caused by the reduction of acquisition time or dose of radiation. There is a general trend towards more specialized segmentation approaches, targeted at a limited set of organs and imaging parameters. Deformable Models are very commonly used and adapt to these requirements e.g. by exploiting prior knowledge, but require a good initialization. Different approaches, including the General Hough Transform and Registration methods will be considered.

  booktitle = {3D Multiscale  Physiological Human },
  author = {Becker, M. and Magnenat-Thalmann, N.},
  title = {Deformable Models in Medical Image Segmentation},
  publisher = {Springer},
  year = {2014},
  topic = {Medical},
  project = {MultiscaleHuman}