| Piecewise linear transformation |
| Image Registration and Fusion Systems |
 | |  |
Given a set of corresponding landmarks in two images, the landmarks in one of the images is triangulated and from the
correspondence between the landmarks, corresponding triangles in the other images are determined. Then, an affine
transformation is determined to map each triangular region in one image to its corresponding triangular region in the other image.
Therefore, unlike a global transformation, this transformation locally registers small corresponding regions in the images.
Because of the same reason, it can register image areas falling inside the convex hull of the landmarks in each image, leaving
areas outside the convex hull of the points unregistered..
A triangulation example is shown below. When images have local geometric differences, this transformation is a better model to
register the images than surface spline.
correspondence between the landmarks, corresponding triangles in the other images are determined. Then, an affine
transformation is determined to map each triangular region in one image to its corresponding triangular region in the other image.
Therefore, unlike a global transformation, this transformation locally registers small corresponding regions in the images.
Because of the same reason, it can register image areas falling inside the convex hull of the landmarks in each image, leaving
areas outside the convex hull of the points unregistered..
A triangulation example is shown below. When images have local geometric differences, this transformation is a better model to
register the images than surface spline.
| A triangulation example. |
To obtain a license for this piecewise linear transformation,
follow this link =>
follow this link =>
