Alignment of contrast enhanced medical images.
Doctoral thesis, UCL (University College London).
The re-alignment of series of medical images in which there are multiple contrast variations is difficult. The reason for this is that the popularmeasures of image similarity used to drive the alignment procedure do not separate the influence of intensity variation due to image feature motion and intensity variation due to feature enhancement. In particular, the appearance of new structure poses problems when it has no representation in the original image. The acquisition of many images over time, such as in dynamic contrast enhanced MRI, requires that many images with different contrast be registered to the same coordinate system, compounding the problem. This thesis addresses these issues, beginning by presenting conditions under which conventional registration fails and proposing a solution in the form of a ’progressive principal component registration’. The algorithm uses a statistical analysis of a series of contrast varying images in order to reduce the influence of contrast-enhancement that would otherwise distort the calculation of the image similarity measures used in image registration. The algorithm is shown to be versatile in that it may be applied to series of images in which contrast variation is due to either temporal contrast enhancement changes, as in dynamic contrast-enhanced MRI or intrinsically in the image selection procedure as in diffusion weighted MRI.
|Title:||Alignment of contrast enhanced medical images|
|Open access status:||An open access version is available from UCL Discovery|
|UCL classification:||UCL > School of BEAMS > Faculty of Engineering Science > Medical Physics and Bioengineering|
Archive Staff Only