FIRST REVIEW REPORT

INTRODUCTION
Magnetic resonance has become the main modality for brain imaging that facilitates safe, noninvasive assessment and monitoring of patients with neurodegenerative diseases such as Parkinsons disease, Alzheimers disease (AD), epilepsy, schizophrenia, and multiple sclerosis (MS). The ability to diagnose and characterize these diseases in vivo using MR image data promises exciting developments both toward understanding the underlying pathologies, as well as conducting clinical trials of drug treatments. One important biomarker that is often used to assess patients with neurode- generative disease is brain tissue volume. The typical rate of global brain atrophy inMS patients has been shown to be 0.6% 0.8% annually, which is two to three times the normal atrophy rate. Evidence has shown that white matter (WM) and gray matter (GM) atrophy at different rates, and each correlates differently to disability thus, accurate measurement of theWMand GMbrain tissues can provide valuable quantitative indicators of disease progression and, potentially, treatment out- comes. Thus, the main goal of this paper is to introduce an automatic algorithm for robust WM, GM, and cerebrospinal uid (CSF) segmentation to facilitate accurate measurement of brain tissues. One popular family of brain tissue segmentation methods is based on normalizing the brain scans by registering (or aligning) them to a pre dened atlas of brain tissues. A second family of brain tissue segmentation methods assigns a label for each tissue based on image statistics either by clustering or by modeling the brain tissue intensity distributions as a nite mixture of distributions such as EM , maximum a posteriori (MAP) , simulated annealing , and Gaussian mixture modeling (GMM) . Other approaches incorporate additional regional information, which is lacking from these statistical methods, into their segmentation frame- work. Such methods extend clustering or EM by integrating with fuzzy connectedness , topological constraints , Gibbs random eld (GRF) , and hidden Markov random eld (HMRF) in the segmentation task. A common difficulty with many of these methods, particularly the random eld approaches, is the requirement for proper parameter settings in a supervised setting.

A third family of brain tissue segmentation methods is based on utilizing geometric information such as deformable models or active contours that delineates region boundaries using a minimization of an energy functional . Deformable models employing level sets provide an effective implicit representation rather than explicit parameterization of the evolving contour. However, a common problem of directly applying the active contour approach in segmenting brain MR images is leakage through weak or noisy edges that are ubiquitous, especially for edge-based deformable models, e.g., geodesic active contour , which describe the evolution of propagating curve as a function of image gradient features. Some researchers incorporated image statistics into their deformable models in various segmentation applications by utilizing coupled surface principle and fuzzy logic , to achieve better stability. Others employed a region-based model by utilizing regional homogeneity in a curve evolution perspective and a hierarchical implementation on brain pathology images. More recently, tissue segmentation was performed and quantitatively evaluated by using the multiphase 3-D level set Segmentation (M3DLS) algorithm.M3DLS utilizes multiphase extension of the region-based deformable model based on the Mum fordShah functional by iteratively deforming two closed curves separating four regions. This minimal partitioning approach assumes piecewise constant or piece-wise smooth data and optimizes a sum of terms, including the lengths and areas for the two closed curves, and the sums of square intensity differences from the means for all four separated regions. This minimization is also performed in a level set framework implicitly. Further extending this model to N- phase allows separation of 2N regions but the number of classes to be segmented is limited to two to the power of the number of closed curves dened. Moreover, complexity increases as more level sets are required and the rate of convergence is typically slow.

BLOCK DIAGRAM

Original Image

Edge Based Segmentation

Image Edge Geometr y

Statistical Homogeneit y

3D Segmentatio n

Segmentation Seg 1 Seg 2

SCOPE OF THE PROJECT

In this paper we proposed a 3-D brain MR segmentation method based on deformable models and demonstrated accurate and stable brain tissue segmentation on single as well as multiple MR sequence scans.

MODULE SEPARATION
MODULE 1: Edge based deformable model. MODULE 2: Proposed hybrid geometric statistical feature. MODULE 3: Segmentation of brain MRI. MODULE 4: Extension to multiple MR.

MODULE DESCRIPTION MODULE 1: EDGE BASED DEFORMABLE MODEL

In first module, we utilize the geodesic active contour model Rather than the region-based formulation due to its Computation soundness and extendibility. The value of

this feature function determines the propagation of the surface by searching for the minimal Riemannian distance. An ideal edge would ultimately have a feature value of zero at all the pixel points along this boundary. This process is subjective and ideal parameters are often difficult to derive for a fully automated segmentation framework.

MODULE 2: PROPOSED HYBRID GEOMETRIC STATISTICAL FEATURE

In this module to transform the feature function g in the traditional geodesic active contour formulation into a hybrid feature function by incorporating geometric image features with voxel statistics to help automate and regularize the evolving contours. This parameter estimation problem for GMM solved by employ EM algorithm to image intensity histogram.

MODULE 3: SEGMENTATION OF BRAIN MRI

To segment the brain tissues, we rst estimate the GMM parameters such that each mixture distribution represents one single class. Based on these estimated distributions, the normalized posterior probability of each voxel is calculated. We derive the hybrid geometricstatistical feature as described above by combining both the voxel statistics and the image gradient information.

MODULE 4: EXTENSION TO MULTIPLE MR

We extend our method so that information from multiple MR sequences with different contrast properties can be incorporated when the data is available. Assuming registered images, we rst replace the geometric feature component in the proposed hybrid active contour feature with the multidimensional vector gradient norm derived from all available data sequences. The resulting probability replaces the posterior probability derived from single contrast input in the segmentation procedure.

ALGORITHM USED

1. Multiple 3-D levels set segmentation. 2. EM algorithm.

FUTURE ENHANCEMENT
Enhance the statistical distribution estimation process by using complex intensity distribution estimation method.

ADVANTAGES
1. Accurate, robust and automated tissue segmentation of brain magnetic resonance image (MRI) data of single as well as multiple MR sequence. 2. Novel hybrid geometric statistical feature to govern contour convergence and extract the complex automatic structure.

APPLICATION
Medical uses for demonstrated consistent and robust result when segmentation MRI scans of both multiple sclerosis and Alzheimers disease patients.

LITERATURE SURVEY [1] M. K. Beyer, C. C. Janvin, J. P. Larsen, and D. Aarsland, An MRI study of patients with Parkinsons disease with mild cognitive impairment and dementia using voxel based morphometry,
We studied regional gray matter density in the hippocampus in Parkinsons disease (PD) patients. We obtained magnetic resonance scans in 44 PD patients (PD patients with dementia (PDD) = 9, non-demented PD patients with visual hallucinations (PD + VH) = 16, and PD patients without dementia and without visual hallucinations (PD - VH) = 19) and 56 controls matched for age and years of education. A region of interest (ROI) of the hippocampus following voxel-based morphometry (VBM) procedures was used to perform group comparisons, single-case individual analysis and correlations with learning scores.

Group comparisons showed that PDD patients and PD+VH patients had significant hippocampal gray matter loss compared to controls. In PDD patients, hippocampal gray matter loss involved the entire hippocampus and in PD+VH this reduction was mainly confined to the hippocampal head. 78 % of PDD patients, 31 % of PD+VH patients and 26 % of PD-VH patients had hippocampal head gray matter loss when compared to controls. These results suggest that in PD the neurodegenerative process in the hippocampus starts in the head of this structure and later spreads to the tail and that, in addition, memory impairment assessed by Reys Auditory Verbal Learning Test (RAVLT) correlates with hippocampal head gray matter loss.

[2] P. E. Grant, StructuralMR imaging, Epilepsia

Summary: Phased-array imaging at 1.5 and 3 T significantly improves image quality compared with that in routine 1.5-T head coil studies. These improvements significantly increase lesion detection in patients with focal epilepsy. Semiautomated image-analysis techniques have the potential to improve lesion detection, assess lesion burden more accurately, characterize cortical abnormalities, and determine the location and extent of associated cortical and deep gray nuclei involvement. With increasing interest in understanding the role white matter plays in epileptogenesis and seizure propagation, diffusion tensor imaging may yield useful information about changes in white matter organization. Our current multimodality imaging combines these structural imaging advances with coregistered neurophysiological information obtained with magnetoencephalography and with simultaneous EEG recordings. Our initial experience suggests that multimodality imaging will improve our ability to detect and define the extent of epileptogenic lesions.

[3] A. Traboulsee, G. Zhao, and D. K. B. Li, Neuroimaging in multiple sclerosis, Neurol. Clin.,
MRI is the dominant neuroimaging modality for multiple sclerosis (MS). Revised diagnostic criteria formally incorporate abnormalities on MRI for diagnosis. MRI is wellsuited for evaluating dynamic changes in MS patients.The suppression of new gadoliniumenhanced T1-weighted and newly active T2-weighted lesions on MRI are now standard outcome measures in clinical trials. This article reviews how conventional and advanced MRIs provide important biomarkers of MS pathology and considers the role of MRI outcomes in clinical trials and in clinical practice.

[4] D. T. Chard, C. M. Griffin, G. J. M. Parker, R. Kapoor, A. J. Thompson, and D. H. Miller, Brain atrophy in clinically early relapsing-remitting multiple sclerosis, Brain,
Brain atrophy measured by MRI is a potentially useful tool for monitoring disease progression in multiple sclerosis. The location, extent and mechanisms of brain atrophy in early disease are not well documented. Using quantitative MRI, this study investigated whole brain, grey and white matter atrophy in clinically early relapsing-remitting multiple sclerosis and its relationship to lesion measures. Data came from 27 normal control subjects (14 females and 13 males, mean age 36.1 years) and 26 subjects with clinically definite multiple sclerosis (18 females and eight males, mean age 35.1 years, mean delay from first symptom to scan 1.8 years, median Expanded Disability Status Scale score 1.0). All had three-dimensional fast spoiled gradient recall (3D FSPGR), T1-weighted pre- and post-gadolinium-enhanced and T2-weighted scans. The 3D FSPGR images were automatically segmented into grey and white matter and cerebrospinal fluid using SPM99. 3D FSPGR hypo-intense, T2 hyper-intense, T1 hypo-intense and T1 post-gadoliniumenhancing lesion volumes were determined by semi-automatic lesion segmentation. The SPM99 output was combined with the 3D FSPGR lesion segmentations to quantify tissue

volumes as fractions of total intracranial volumes, producing values for the brain parenchymal fraction (BPF), white matter fraction (WMF) and grey matter fraction (GMF). Comparing multiple sclerosis with control subjects, BPF, GMF and WMF were significantly reduced (P < 0.001 for all tissue fractions). Using Pearson correlations, T2 hyper-intense and T1 hypo-intense lesion volumes were inversely related to BPF (T2 r = -0.78, P < 0.001; T1 r = -0.59, P = 0.002) and GMF (T2 r = -0.73, P < 0.001; T1 r = -0.53, P = 0.006), but not WMF (T2 r = -0.30, P = 0.134; T1 r = -0.26, P = 0.199). T1 postgadolinium-enhancing lesion volumes were not correlated with any fractional volumes. These results indicate that significant brain atrophy, affecting both grey and white matter, occurs early in the clinical course of multiple sclerosis. The lack of correlation between lesion load measures and WMF suggests that pathological changes in white matter may occur by mechanisms which are at least partly independent from overt lesion genesis in early multiple sclerosis.

[5] A. W. C. Liew and Y. Hong, Current methods in automatic tissue segmentation of 3D magnetic resonance brain images, Current Med. Imag. Rev.,
To improve the efficiency of brain image analysis, we propose a full-automatic method for extracting brain tissue from three-dimensional magnetic resonance imaging of T1-weighted data on the human head (brain tissue extraction method using erosion-dilation treatment [BREED]). The extraction processing is realized by combining signal intensity thresholding by means of the discriminant analysis method and an erosion-dilation treatment of the image. The accuracy of BREED is evaluated using both simulated and subject data. BREED can extract brain tissues with high accuracy (approximately 97%) for either simulated or subject data.

[6] D. L. Pham and J. L. Prince, Adaptive fuzzy segmentation of magnetic resonance images, IEEE Trans. Med. Imag.,
A new approach for robust segmentation of magnetic resonance images is described. The approach is derived from a generalization of the objective function used in Pham and Prince's Adaptive Fuzzy C-means algorithm (AFCM). Within the objective

function, an additional constraint is placed on the membership functions that forces them to be spatially smooth. Minimization of this objective function results in an unsupervised fuzzy segmentation algorithm that is robust to both intensity inhomogeniety artifacts as well as noise and other artifacts. The efficacy of the algorithm is demonstrated on simulated magnetic resonance images.

[7] M.W.Wells,W. E. L. Grimson, and F. A. Jolesz, Adaptive segmentation of MRI data, IEEE Trans. Med. Imag.,
Intensity-based classification of MR images has proven problematic, even when advanced techniques are used. Intrascan and interscan intensity inhomogeneities are a common source of difficulty. While reported methods have had some success in correcting intrascan inhomogeneities, such methods require supervision for the individual scan. This paper describes a new method called adaptive segmentation that uses knowledge of tissue intensity properties and intensity inhomogeneities to correct and segment MR images. Use of the expectation-maximization (EM) algorithm leads to a method that allows for more accurate segmentation of tissue types as well as better visualization of magnetic resonance imaging (MRI) data, that has proven to be effective in a study that includes more than 1000 brain scans. Implementation and results are described for segmenting the brain in the following types of images: axial (dual-echo spin-echo), coronal [three dimensional Fourier transform (3-DFT) gradient-echo T1-weighted] all using a conventional head coil, and a sagittal section acquired using a surface coil. The accuracy of adaptive segmentation was found to be comparable with manual segmentation, and closer to manual segmentation than supervised multivariant classification while segmenting gray and white matter.

REFERENCES
[1] M. K. Beyer, C. C. Janvin, J. P. Larsen, and D. Aarsland, An MRI study of patients with Parkinsons disease with mild cognitive impairment and dementia using voxel based morphometry, J. Neurol. Neurosurg. Psychiatry, vol. 78, no. 3, pp. 254259, Mar. 2007. [2] M. Grossman, C.McMillan, P.Moore, L. Ding, G. Glosser, M.Work, and J. Gee, Whats in a name: Voxel-based morphometry analysis of MRI and naming difficulty in Alzheimers disease, frontotemporal dementia and corticobasal degeneration, Brain, vol. 127, no. 3, pp. 628649, 2004.

[3] P. E. Grant, StructuralMR imaging, Epilepsia, vol. 45, no. s4, pp. 416, 2004. [4] J. J. Wisco, G. Kuperberg, D. Manoach, B. T. Quinn, E. Busa, B. Fischl, S. Heckers, and A. G. Sorensen, Abnormal cortical folding patterns within Brocas area in schizophrenia: Evidence from structural MRI, Schizophrenia Res., vol. 94, no. 13, pp. 317327, Aug. 2007. [5] D. W. Paty, D. Li, and G. J. Zhao,MRI in Multiple Sclerosis-Implications for Diagnosis and Treatment, 2nd ed. UBC MS/MRI Research Lab., Vancouver, BC, Canada, Rep. for Ares-Serono SA, 1999. [6] D. H. Miller, Biomarkers and surrogate outcomes in neurodegenerative disease: Lessons from multiple sclerosis, NeuroRx, vol. 1, pp. 284294, 2004. [7] A. Traboulsee, G. Zhao, and D. K. B. Li, Neuroimaging in multiple sclerosis, Neurol. Clin., vol. 23, pp. 131148, 2005. [8] D. T. Chard, C. M. Griffin, G. J. M. Parker, R. Kapoor, A. J. Thompson, and D. H. Miller, Brain atrophy in clinically early relapsing-remitting multiple sclerosis, Brain, vol. 125, no. 2, pp. 327337, Feb. 2002. [9] N. De Stefano, P. M. Matthews, M. Filippi, F. Agosta, M. De Luca, M. L. Bartolozzi, L. Guidi, A. Ghezzi, E. Montanari, A. Cifelli, A. Federico, and S. M. Smith, Evidence of early cortical atrophy in MS, Neurology, vol. 60, pp. 11571162, 2003. [10] J. Sastre-Garriga,G.T. Ingle, D.T. Chard, L. Ramio-Torrenta, D. H. Miller, and A. J. Thompson, Grey and white matter atrophy in early clinical stages of primary progressive multiple sclerosis, NeuroImage, vol. 22, no. 1, pp. 353359, May 2004. [11] V. M. Anderson, N. C. Fox, and D. H. Miller, Magnetic resonance imaging measures of brain atrophy in multiple sclerosis, J. Magn. Reson. Imag., vol. 23, pp. 605 618, 2006. [12] S. K. Warfield, K. H. Zou, and W. M. Wells, Simultaneous truth and performance level estimation (STAPLE): An algorithm for the validation of image segmentation, IEEE Trans.Med. Imag., vol. 23, no. 7, pp. 903 921, Jul. 2004. [13] D. L. Pham, C. Xu, and J. L. Prince, Current methods in medical image segmentation, Annu. Rev. Biomed. Eng., vol. 2, pp. 315337, Aug. 2000. [14] K. M. Pohl, R. Kikinis, and W. M. Wells, Active mean fields: Solving the mean field approximation in the level set framework, in Proc. IPMI 2007, pp. 2637. [15] M. Rousson, N. Paragios, and R. Deriche, Implicit active shape models for 3D segmentation in MR imaging, in Proc. MICCAI 2004, pp. 209 216.

[16] L. Ibanez, W. Schroeder, L. Ng, and J. Cates, The ITK Software Guide, 2nd ed. New York: Kitware, Inc., 2005. [17] A. P. Dempster, N. M. Laird, and D. B. Rubin, Maximum likelihood from incomplete data via the EM algorithm, J. R. Stat. Soc. Ser. B (Methodological), vol. 39, no. 1, pp. 1 38, 1977. [18] A. Huang, R. Abugharbieh, R. Tam, and A. Traboulsee, Automatic MRI brain tissue segmentation using a hybrid statistical and geometric model, in Proc. IEEE ISBI, 2006, pp. 394397. [19] R. C. Gonzales and R. E. Woods, Digital Image Processing, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall, 2002, pp. 549.