Abstract: Multiphase level set model is sensitive to initial contour curve and has huge computation in
the process of the multiple objectsʹ segmentation. A novel Image segmentation method is presented
for multiphase scenario, which initializes the multiphase level set function by coarse image
segmentation using fuzzy C-means clustering algorithm and applies graph cuts algorithm to acquire
multiphase output image. The method can effectively reduce the sensitivity of the multiphase level set
algorithm to initial contour and is easier to gain the multiphase output image by graph cuts algorithm.
At the same time, the multiphase level set function quickly converge to the minimum energy value
with small amount of calculation and high computational efficiency using the graph cuts algorithm.
The experiments show that this method has better segmentation effect and higher efficiency of image
segmentation.

Key words: fuzzy C-means clustering, image segmentation, graph cuts, multiphase level set