Two‐stage multishape segmentation of brain structures using image intensity, tissue type, and location informationa)

Abstract

Purpose:The authors propose a fast, robust, nonparametric, entropy-based, coupled, multishape approach to segment subcortical brain structures from magnetic resonance images ͑MRIs͒. Methods: The proposed method uses three types of information: Image intensity, tissue types, and locations of structures. The image intensity information is captured by estimating the probability density function ͑pdf͒ of the image intensities in each structure. The tissue type information is captured by applying an unsupervised tissue segmentation method to the image and estimating a probability mass function ͑pmf͒ for the tissue type of each structure. The location information is captured by estimating pdf of the location of each structure from the training datasets. The resulting pmf's and pdf's are used to define an entropy function whose minimum corresponds to a desirable segmentation of the structures. The authors propose a three-step optimization strategy for the segmentation method. In the first step, a powerful automatic initialization method is developed based on tissue type and location information of the structures. In the second step, a quasi-Newton method is used to optimize the parameters of the energy function. To speed up the iterations, derivatives of the energy function with respect to its parameters are analytically derived and used in the optimization process. In the last step, the limitations related to the prior shape model are removed and a level-set method is applied for the fine tuning of the segmentation results. Results: The proposed method is applied to two different datasets and the results are compared to those of previous methods in literature. Experimental results are presented for lateral ventricles, caudate, thalamus, putamen, pallidum, hippocampus, and amygdala. Conclusions:The results illustrate superior performance of the proposed segmentation method compared to other methods in literature. The execution time of the algorithm is a few minutes, suitable for a variety of applications.

Multi-atlas multi-shape segmentation of fetal brain MRI for volumetric and morphometric analysis of ventriculomegaly

“…We formulate the problem in a probabilistic framework and solve the optimization problem through a numerical approach, in which the initial solution is perturbed and deformed to minimize the energy functional. This approach is based on the recently developed probabilistic shape optimization framework presented in Akhondi-Asl and Soltanian-Zadeh (2010). For a literature review of the previous multi-shape segmentation methods we refer to the recent papers by Akhondi-Asl and Soltanian-Zadeh (2009, 2010).…”

Cortical brain structures segmentation using constrained optimization and intensity coupling

“…The method represents a modified version of our previous work described in [5][6]. Specifically, we have modified the pdf estimation of the intensity of the structures to consider the intensity relationships among similar structures.…”

“…This is different from the tk fk sk p x p x p x in [5], which models them as independent pdf and pmf. The new probability mass function is more accurate, because it considers the relationship between the intensity and location of the structures.…”

“…In considering these relationships, we defined a pmf that takes into account both location and tissue type information. First, we clustered the intensity image I(x) based on the method in [5] to 10 classes to extract the image S(x). In addition, we clustered, 1,..., S x are in the same class; further, it will be smaller than one based on the difference between the two classes.…”

“…In [4] we have added constraint to the shape parameters to achieve a more robust segmentation algorithm. In [5] we have modified the energy function by considering tissue type and location of the structures as independent information. However, it should be noted that the model of tissue type was for limited number of structures.…”