Automatic Brain Tumor Segmentation with Normalized Gaussian Bayesian Classifier and Fluid Vector Flow

• Author(s) / Creator(s)

  • Wang, Tao

• Technical report TR09-11. An automatic brain tumor segmentation method is presented in this paper. This method has 3 stages. In the first stage, a so-called Normalized Gaussian Mixture Model (NGMM) is proposed and used to model the brain tissues. In the second stage, a Gaussian Bayesian Classifier based on the NGMM and the prior probabilities of different brain tissues is exploited to acquire a so-called Gaussian Bayesian Brain Map (GBBM) from the test 3D brain MR images. GBBM is further processed to highlight the brain tumor and initialize a so-called Fluid Vector Flow (FVF) algorithm. In the last stage, FVF is used to segment the brain tumor. The major contribution of this paper is two-fold. First, we present a NGMM to represent healthy brains. This model can be easily modified for modeling other tasks in various application domains. Second, we extend our 2D FVF algorithm to 3D space and use it for automatic brain tumor segmentation. This method has been validated on a publicly available dataset containing 10 T1 Magnetic Resonance (MR) images of 3 types of brain tumor. The results demonstrate that this technique can automatically generate 3D segmentation images of multiple types of brain tumor solely from T1 MRIs. | TRID-ID TR09-11

• Date created

  2009

• Subjects / Keywords

  • 3D imaging
  • Automatic brain tumor segmentation
  • Normalized Gaussian Mixture Model

• Type of Item

  Report

• DOI

  https://doi.org/10.7939/R39C6S153

• License

  Attribution 3.0 International