Authors : Akbar Alipour Sifar and Mousa Shamsi
Abstract: Segmentation of brain MR images is a major issue in medical image processing computations. In brain MR images, segmentation is caused by an inherent error which is called intensity in homogeneity. This is due to the existence of an overlap between different brain tissues which often causes false classification of tissues. This paper uses a new proposed method for segmentation and bias field correction simultaneously which is called Multiplicative Intrinsic Component Optimization (MICO). The proposed method, breaks down MR images into two components, one component characterizes a physical property of tissue and other inherent bias field that accounts for the intensity in homogeneity with spatial features. Then, via. energy minimization in an iterative process, the above components are optimized and consequently, segmentation and bias field correction was carried out, simultaneously. Qualitative assessment of MICO method was proved in terms of accuracy and robustness and showed high accuracy of about 90% for bias field correction and segmentation in three areas of the brain, especially in the area containing the Cerebrospinal Fluid (CSF).
Akbar Alipour Sifar and Mousa Shamsi, 2019. The Quantitative and Qualitative Evaluation of Simultaneous Segmentation using Multiplicative Intrinsic Component Optimization in Brain MR Images. International Journal of Signal System Control and Engineering Application, 12: 143-147.