Multimodal Medical Image Registration Based on an Information-Theory Measure with Histogram Estimation of Continuous Image Representation

Algorithm 1

Fast continuous histogram estimation.

Input: Two images and .

Stage 1: The coordinates of sample points of and are obtained by random sampling strategy.

Stage 2: Each sample point is partitioned into two triangles (2D) or six tetrahedrons (3D case).

Stage 3: The isointensity lines or isosurfaces of the corresponding sample points from two images are constructed.

Stage 4: The joint histogram between and is updated using the isointensity lines or isosurfaces. Go to stage 2,

unless all sample points are exhausted.

Output: The joint histogram between and .