3D Liver Tumor Segmentation in CT Images Using Improved Fuzzy C-Means and Graph Cuts

Algorithm 2

Graph cuts segmentation.

input: the tumor VOI , the gradient magnitude , the probabilistic models

for tumor and liver

output: the binary tumor mask

Set parameters , ,

Generate initial seed mask. Set inside the confidence connected

segmented tumor regions (), and outside the liver region or

adjacent to the boundaries of ()

Compute the mean value of . Set , and then generate sigmoid

of according to (1)

Construct the graph on

for   on ,   do

Compute region cost according to (9)

Compute boundary cost according to (10), using the 6-neighborhood

system

, if

Find optimal tumor surface via the max-flow/min-cut algorithm. Set

for and for