Abstract
Liver tumors segmentation is an important prerequisite for planning of surgical interventions. For clinical applicability, the segmentation approach must be able to cope with the high variation in shape and gray-value appearance of the liver. We present a fully automatic 3D segmentation method for the liver tumors from contrast-enhanced CT data. The method consists of two main stages. First an initial histogram and statistical distribution functions are created, and from them a new image is created where, in each voxel, a weighted function is attached in accordance with the probability of the voxel grey level. Next, we use the active contour method on the new image, where the active contour evolution is based upon the minimization of variances between the liver tumor and its closest neighborhood.
Keywords
Source Code and Data
No source code files available for this publication.
Reviews
Xiang Deng
Friday 25 July 2008
This paper presents an automatic live tumor segmentation technique.
In this method, probalistic intensity models of the tumor and surrounding liver tissue are first estimated using a non-parametric technique. Second, the tumors are segmented using a 3D active contour model that can minimize the intensity variance inside and outside the object of interest.
Detailed comments:
1) The active surface model used in the segmentation seems to require some kind of initialization, e.g., manually defined initial contour. If that's the case, the method should be semi-automatic. Could the authors clarify this point?
2) The meaning of symbols in equation (1) in section 2.1 are not clearly defined, i.e., does "measure of the region of the tumor" mean the volume of the tumor? what does |Vol_{in}| stand for?
3) In "Experimental results" section, the statement "the small relative difference between the intensity of the tumor and the healthy part this cause that any energy based segmentation of the normalized intensity values will not work" is too strong.
4) In "Experimental results" section, I don't see a clear advatange of the proposed distribution estimation method from Figures 2 and 3. Could the authors show some quantitative comparison?