MRI segmentation: Methods and applications

A statistical model is presented that represents the distributions of major tissue classes in single-channel magnetic resonance (MR) cerebral images. Using the model, cerebral images are segmented into gray matter, white matter, and cerebrospinal fluid (CSF). The model accounts for random noise, magnetic field inhomogeneities, and biological variations of the tissues. Intensity measurements are modeled by a finite Gaussian mixture. Smoothness and piecewise contiguous nature of the tissue regions are modeled by a three-dimensional (3-D) Markov random field (MRF). A segmentation algorithm, based on the statistical model, approximately finds the maximum a posteriori (MAP) estimation of the segmentation and estimates the model parameters from the image data. The proposed scheme for segmentation is based on the iterative conditional modes (ICM) algorithm in which measurement model parameters are estimated using local information at each site, and the prior model parameters are estimated using the segmentation after each cycle of iterations. Application of the algorithm to a sample of clinical MR brain scans, comparisons of the algorithm with other statistical methods, and a validation study with a phantom are presented. The algorithm constitutes a significant step toward a complete data driven unsupervised approach to segmentation of MR images in the presence of the random noise and intensity inhomogeneities.

Statistical approach to segmentation of single-channel cerebral MR images

A new region growing method for finding the boundaries of blobs is presented. A unique feature of the method is that at each step, at most one pixel exhibits the required properties to join the region. The method uses two novel discontinuity measures, average contrast and peripheral contrast, to control the growing process.

Region growing: a new approach

A robust fully automatic method for segmenting the brain from head magnetic resonance (MR) images has been developed, which works even in the presence of radio frequency (RF) inhomogeneities. It has been successful in segmenting the brain in every slice from head images acquired from several different MRI scanners, using different-resolution images and different echo sequences. The method uses an integrated approach which employs image processing techniques based on anisotropic filters and "snakes" contouring techniques, and a priori knowledge, which is used to remove the eyes, which are tricky to remove based on image intensity alone. It is a multistage process, involving first removal of the background noise leaving a head mask, then finding a rough outline of the brain, then refinement of the rough brain outline to a final mask. The paper describes the main features of the method, and gives results for some brain studies.

/pdf/fully-automatic-segmentation-of-the-brain-in-mri-4fnd6after.pdf

Fully automatic segmentation of the brain in MRI

Brain image segmentation is one of the most important parts of clinical diagnostic tools. Brain images mostly contain noise, inhomogeneity and sometimes deviation. Therefore, accurate segmentation of brain images is a very difficult task. However, the process of accurate segmentation of these images is very important and crucial for a correct diagnosis by clinical tools. We presented a review of the methods used in brain segmentation. The review covers imaging modalities, magnetic resonance imaging and methods for noise reduction, inhomogeneity correction and segmentation. We conclude with a discussion on the trend of future research in brain segmentation.

https://link.springer.com/content/pdf/10.1007%2Fs10462-010-9155-0.pdf

Review of brain MRI image segmentation methods

Automatic segmentation of head MRI images by knowledge guided thresholding.

Up to now there have been reports on a number of comparative evaluations on the capabilities of region segmentation methods which are one part of image segmentation. However, evaluations on the ease of setting the required parameters and their stabilities when these methods are used have not been performed yet. Therefore, in this paper, we will quantitatively stuff the relationships between the parameters and the characteristics (shape and orientation) of the target objects in the image for several representative region segmentation methods, and examine the accuracy of the extraction and parameter stability. The methods dealt with are (1) iterative merging method, (2) a method based on statistical hypothesis test, (3) a heuristic method, (4) a split and merge method, (5) a relative similarity method, and (6) a region- and edge-based method

Comparative performance evaluation of segmentation methods based on region growing and division

This paper proposes a segmentation method based on region information and edge information. The method consists of three steps: (1) a connected component which is a part of an object is extracted as a seed by using a simple process such as a histogram analysis; (2) this seed is expanded repeatedly; and (3) to evaluate the possibility of the expansion, the uniformity of gray levels (region information) and checking the presence of border pixels (edge information) are used.



Although the method can be considered as a kind of region expansion, the significant features are that it contains the growth of a seed and the edge information (a part of control information) to expand the region repeatedly. The former gives preliminary information on approximate gray levels of an object and extracts the object as a connected component, and the latter makes the edge information more useful. Although the usefulness of the edge information in a conventional region-and-edge integrated method such as the superslice algorithm depends greatly on its region segmentation, the use of such methods can be avoided in the proposed method, since both the region and edge information are used simultaneously for the control conditions of the region expansion. The method is justified experimentally by using synthetic images and actual images in two and three dimensions.

A segmentation method based on region information and edge information

Automatic segmentation of mri head images by 3-D region growing method which utilizes edge information

This paper discusses the structural analysis algorithm for line patterns, based on the distance transformation of a line pattern by iterative local parallel processing. Although parallel processors have been applied recently to image processing, the processing of line patterns in the past was based primarily on the transformation of the object pattern into the list structure or symbol strings. Few studies have been made on the parallel processing algorithm for a line pattern. From such a viewpoint, this paper presents parallel-type algorithms for loop counting, loop extraction and arc decomposition, among the procedures for processing line patterns. Properties of these algorithms also are discussed. The method proposed here is based on the distance transformation (Min. type) for line patterns we developed, and can be performed by iterative local parallel processing, in the same way as the distance transformation itself.

Hidetomo Suzuki

Papers

Automatic segmentation of head MRI images by knowledge guided thresholding.

Comparative performance evaluation of segmentation methods based on region growing and division

A segmentation method based on region information and edge information

Automatic segmentation of mri head images by 3-D region growing method which utilizes edge information

Structural analysis of line patterns using distance transformation (DTLP)