Tumor Manual segmentation of these regions from huge data

Tumor is an uncontrolled growth of
unnatural cells in any part of the body 38. Tumor affects the brain tissues
and makes unpredictable abnormal cell development is called as brain tumor. Early
detection and diagnosis of brain tumor is very much important to improving the
treatment planning and increases survival rate of the patients 48. Generally
brain tumor classified into primary brain tumors and malignant brain tumors. Primary
brain tumor start and spread into brain itself whereas malignant initial origin
maybe anywhere in the body and start to spread inside the brain.

The most common primary brain tumor is
called Gliomas, which originate from Glial cells and infiltrate the surrounding
tissues. Majority of brain tumor segmentation methods are focus on Gliomas
because of Gliomas are mostly targets the adults. Segmentation of Gliomas is
very difficult task due to variety of shapes, sizes and any position in the
brain lecture note 75 page. MRI is non-invasive techniques provide a good
contrast on soft tissues using radio frequency and powerful magnetic field. MRI
generates stack of 2D slices in multisequence format based on altering
excitation and repetition time. The multisequence used for the Gliomas
diagnosis are: T1-weighted (T1), T1-weighted with gadolinium contrast enhancement
(T1c), T2-weighted (T2) and Fluid Attenuated Inversion Recovery (FLAIR) 46.
Sample multisequence images are shown in Fig.1.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Brain tumor segmentation consists of
segmenting active tumor, edema, and necrosis regions from normal tissues such
as gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) and shown
in Fig.1. Manual segmentation of these regions from huge data is a time
consuming painstaking task to the clinicians and prone to substantial
inter-rater error. To remove all these difficulties, automatic methods and
algorithms for classification, segmentation and visualization along with the
modern systems provided a robust, repeatable, accurate and fast segmentation
39.

    

 

 

                       (a)                                (b)

 

 

                                                                  

 

                      (c)                                 (d)                                  (e)              

Fig. 1 MRI multisequence
and tumor substructures (a) T1-weighted (b) T1c (c) FLAIR (d) T2-weighted (e)
Tumor with substructures

The abnormal image selection is a pre-process
of tumor segmentation from the stack of 2D slice. Therefore efficient algorithm
for tumor detection is always demand in the field of classification which helps
to concentrate the diagnosis process with abnormal slice alone 40. In recent
years, feature extraction with support vector machine (SVM), artificial neural
network (ANN), k-nearest neighbor (KNN), random forest (RF), self-organizing
maps (SOM) and deep leaning neural network (DNN) classifiers and its variants
has been proposed for MR brain tumor detection 41 – 44, 56. Fully automatic
methods for segmenting brain tumor from abnormal image are very helpful in the clinical
environment. In past few decades, medical imaging with following soft computing
techniques plays a important role in brain tumor segmentation to archive high
accuracy: Convolutional neural network (CNN), Local ternary pattern (LTP),
Fuzzy computing and genetic algorithm4547 4950.

Segmentation
algorithms need more computational power for large dataset to produce fast and
accurate results 51. Graphics processing unit (GPU) gives the solutions to
these computation problems for large datasets using their features such as,
high computation throughput, high memory bandwidth, support for floating-point
arithmetic and low cost. Compute unified device architecture (CUDA) is a GPU
programming model introduced by NVIDIA Corporation to solve large computational
problems 55. GPU based brain tumor segmentation methods have proposed in
recent years to achieve high computational throughput 52 -54.

In this paper, proposed a fully automatic
brain tumor detection and segmentation method using multimodal Brain Tumor
Segmentation (BraTS) training datasets. We categories proposed method into
three phases: (1) abnormality detection, (2) approximate tumor extraction and
(3) tumor substructures segmentation. For the pre-processing, abnormal slices
were efficiently detected from dataset using feature blocks and SVM classifier.
Approximate tumor region reached using fuzzy c means (FCM) algorithm. A novel probabilistic
ternary pattern (PLTP) is proposed for tumor substructures segmentation from
approximate tumor region. For the post processing, segmented substructures used
to estimate the tumor volume and 3D visualization. Finally, availability of GPU-CUDA
makes the inherent parallelism of FCM algorithm in the proposed method to
reduce the computation time. We report promising results obtained using BraTS
2013 and BraTS 2015 training dataset.

The rest of this paper is structures as
follows: in Section 2 describe the available prior work related with this
experiment; Section 3 explained materials and methods used for the experiment;
Section 4 illustrate the implementation details of proposed methodology; evaluation parameters are discussed in Section 5;
Results and analysis are given in the Section 6; Section 7 concluded the work. 

Tumor is an uncontrolled growth of
unnatural cells in any part of the body 38. Tumor affects the brain tissues
and makes unpredictable abnormal cell development is called as brain tumor. Early
detection and diagnosis of brain tumor is very much important to improving the
treatment planning and increases survival rate of the patients 48. Generally
brain tumor classified into primary brain tumors and malignant brain tumors. Primary
brain tumor start and spread into brain itself whereas malignant initial origin
maybe anywhere in the body and start to spread inside the brain.

The most common primary brain tumor is
called Gliomas, which originate from Glial cells and infiltrate the surrounding
tissues. Majority of brain tumor segmentation methods are focus on Gliomas
because of Gliomas are mostly targets the adults. Segmentation of Gliomas is
very difficult task due to variety of shapes, sizes and any position in the
brain lecture note 75 page. MRI is non-invasive techniques provide a good
contrast on soft tissues using radio frequency and powerful magnetic field. MRI
generates stack of 2D slices in multisequence format based on altering
excitation and repetition time. The multisequence used for the Gliomas
diagnosis are: T1-weighted (T1), T1-weighted with gadolinium contrast enhancement
(T1c), T2-weighted (T2) and Fluid Attenuated Inversion Recovery (FLAIR) 46.
Sample multisequence images are shown in Fig.1.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Brain tumor segmentation consists of
segmenting active tumor, edema, and necrosis regions from normal tissues such
as gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) and shown
in Fig.1. Manual segmentation of these regions from huge data is a time
consuming painstaking task to the clinicians and prone to substantial
inter-rater error. To remove all these difficulties, automatic methods and
algorithms for classification, segmentation and visualization along with the
modern systems provided a robust, repeatable, accurate and fast segmentation
39.

    

 

 

                       (a)                                (b)

 

 

                                                                  

 

                      (c)                                 (d)                                  (e)              

Fig. 1 MRI multisequence
and tumor substructures (a) T1-weighted (b) T1c (c) FLAIR (d) T2-weighted (e)
Tumor with substructures

The abnormal image selection is a pre-process
of tumor segmentation from the stack of 2D slice. Therefore efficient algorithm
for tumor detection is always demand in the field of classification which helps
to concentrate the diagnosis process with abnormal slice alone 40. In recent
years, feature extraction with support vector machine (SVM), artificial neural
network (ANN), k-nearest neighbor (KNN), random forest (RF), self-organizing
maps (SOM) and deep leaning neural network (DNN) classifiers and its variants
has been proposed for MR brain tumor detection 41 – 44, 56. Fully automatic
methods for segmenting brain tumor from abnormal image are very helpful in the clinical
environment. In past few decades, medical imaging with following soft computing
techniques plays a important role in brain tumor segmentation to archive high
accuracy: Convolutional neural network (CNN), Local ternary pattern (LTP),
Fuzzy computing and genetic algorithm4547 4950.

Segmentation
algorithms need more computational power for large dataset to produce fast and
accurate results 51. Graphics processing unit (GPU) gives the solutions to
these computation problems for large datasets using their features such as,
high computation throughput, high memory bandwidth, support for floating-point
arithmetic and low cost. Compute unified device architecture (CUDA) is a GPU
programming model introduced by NVIDIA Corporation to solve large computational
problems 55. GPU based brain tumor segmentation methods have proposed in
recent years to achieve high computational throughput 52 -54.

In this paper, proposed a fully automatic
brain tumor detection and segmentation method using multimodal Brain Tumor
Segmentation (BraTS) training datasets. We categories proposed method into
three phases: (1) abnormality detection, (2) approximate tumor extraction and
(3) tumor substructures segmentation. For the pre-processing, abnormal slices
were efficiently detected from dataset using feature blocks and SVM classifier.
Approximate tumor region reached using fuzzy c means (FCM) algorithm. A novel probabilistic
ternary pattern (PLTP) is proposed for tumor substructures segmentation from
approximate tumor region. For the post processing, segmented substructures used
to estimate the tumor volume and 3D visualization. Finally, availability of GPU-CUDA
makes the inherent parallelism of FCM algorithm in the proposed method to
reduce the computation time. We report promising results obtained using BraTS
2013 and BraTS 2015 training dataset.

The rest of this paper is structures as
follows: in Section 2 describe the available prior work related with this
experiment; Section 3 explained materials and methods used for the experiment;
Section 4 illustrate the implementation details of proposed methodology; evaluation parameters are discussed in Section 5;
Results and analysis are given in the Section 6; Section 7 concluded the work. 

x

Hi!
I'm Elaine!

Would you like to get a custom essay? How about receiving a customized one?

Check it out