AUTOMATED DIAGNOSIS FOR BRAIN TUMOR BASED ON MRI IMAGES PROJECT

  Fig-1: BRAIN (MRI) Key Words MRI (Magnetic resonance imaging) Benign (Benign tumors) Malignant (Malignant tumors) Primary Big Data Brain Tumor Stem Cells Clinical Trial Translational
The post AUTOMATED DIAGNOSIS FOR BRAIN TUMOR BASED ON MRI IMAGES PROJECT first appeared on COMPLIANT PAPERS.

Fig-1: BRAIN (MRI)

Key Words
MRI (Magnetic resonance imaging)
Benign (Benign tumors)
Malignant (Malignant tumors)
Primary
Big Data
Brain Tumor Stem Cells
Clinical Trial
Translational research
Computed Tomography (CT)
Feed Forward Neural Networks (FFNNs)
Convolutional Neural Networks (CNNs)
Recurrent Neural Networks (RNNs)
Generative Adversarial Networks (GANs)
Deep Reinforcement Learning (Deep RL)

Table 1:- key words.

CHAPTER 1
1.1 Introduction
Brain tumors can be classified into two types: benign (noncancerous) and malignant (cancerous). The malignant tumors can quickly spread to other tissues in the brain and lead to worsening the patient’s condition. When most of the cells are old or damaged, they are destroyed and replaced by new cells. if damaged and old cells are not eliminated with generating the new cells, it can cause problems.

The production of additional cells often results in the formation of a mass of tissue, which refers to the growth or tumor. Brain tumor detection is very complicated and difficult due to the size, shape, location, and type of tumor in the brain. Diagnosis of brain tumors in the early stages of the tumor’s start is difficult because it cannot accurately measure the size and resolution of the tumor. the treatment of tumor depends on the timely diagnosis of the tumor. The diagnosis is usually done by a medical examination.

MRI imaging is a method that provides accurate images of the brain and is one of the most common and important methods for diagnosing and evaluating the patient’s brain. Also, can be used to measure the tumor’s size. MRI images provide better results than other imaging techniques such as Computed Tomography (CT).

1.2 Problem Statement & Significance
Brain tumors is a mass of abnormal cells that grows inside or around the brain and multiplies uncontrollably. These cells can be differentiated from the surrounding tissue by their structure. Brain tumors originate in the brain and rarely spread to other sections of the body. However, secondary or metastatic brain tumors are produced from cancer cells in another part of the body that pass to the brain through the bloodstream, in a process which is called metastasis.
Brain tumors represent among the most common human diseases. It has been estimated that the prevalence of brain tumors among the Saudi population is 0.3%. In the Middle East, Iran ranks highest with respect to the highest prevalence rate of brain tumors, with the Kingdom of Saudi Arabia (KSA) ranking second.
The medical imaging plays an important role in the detection of brain tumors because it is non-invasive method, presents different texture features of different tissues and it does not need to surgery. Therefore, we cannot detect the brain tumor without medical imaging.
But human detection of the visible features of the image is limited, which increases the risk of human error in manual segmentation. Therefore, some kind of automated segmentation will always be helpful, especially for large MRI datasets.

1.3 Proposed Solution (System)

â–  Developing a tool that takes images of the brain that can detects the tumors in the simplest way possible with less invasive techniques and time consuming.
â–  List your aims & goals:
● Learning how to develop with a new programming language and experience with the AI.
● Our goal is to develop a tool that works seamlessly with accurate result for the patients and doctors.

1.4 Project Domain & Limitations

1.5 Definitions of New Terms

MRI -Magnetic resonance imaging – A medical imaging technique that uses powerful magnetic fields to make detailed pictures of the inside of the body.
Benign -The least aggressive type of brain tumor is often called a benign brain tumor. They originate from cells within or surrounding the brain, do not contain cancer cells, grow slowly, and typically have clear borders that do not spread into other tissue. They may become quite large before causing any symptoms. If these tumors can be removed entirely, they tend not to return. Still, they can cause significant neurological symptoms depending on their size, and location near other structures in the brain. Some benign tumors can progress to become malignant.
Malignant – Malignant tumors are cancerous. The cells can grow and spread to other parts of the body. Primary – Whether cancerous or benign, tumors that start in cells of the brain are called primary brain tumors. Primary brain tumors may spread to other parts of the brain or to the spine, but rarely to other organs.
Big Data -As biomedical tools and technologies rapidly improve, researchers are producing and analyzing an ever-expanding amount of complex biological data sets that combine vast amounts of information on tumors; this is known as ‘Big Data.’
Brain Tumor Stem Cells – A small population of cells within the brain that have the ability to multiply and self-renew and to differentiate into several types of mature cells. Research is focused on identifying which cells within a brain tumor are the actual stem cells and determining their characteristics. Understanding brain tumor stem cells may lead to approaches to kill them, prevent re-growth or recurrence of brain tumors. Clinical Trial – A type of research study that tests how well new medical approaches work in people. These studies test new methods of screening, prevention, diagnosis, or treatment of a disease. Also called clinical study.
Translational research – A term used to describe the process by which the results of research done in the laboratory are used to develop new ways to diagnose and treat disease.
CHAPTER 2
2.1 Background Information:
Doctors use many tests to find or diagnose a brain tumor. They also do tests to find if it has spread to another part of the body from where it started. Magnetic resonance imaging (MRI) which creates an image of the brain structure. It uses a strong magnetic field to align spinning atomic nuclei within body tissues, then distributes the axis of rotation of these nuclei and observe the radio frequency signal. MRI scans are noninvasive, pose a little health risk. Its disadvantage is that the patient has to hold still for long periods of time in a noisy, cramped space while the imaging is performed.

The formation of abnormal groups of cells inside the brain or near it leads to the initialization of a brain tumor. The abnormal cells abrupt the processing of the brain and affect the health of a patient [1].

Brain tumor is broadly classified into two types: cancerous tumors, known as malignant tumors, and noncancerous tumors, known as benign tumors. Malignant tumors are further classified into grades I to IV and neoplasms more than 120 histological types of these tumors have been by the World Health Organization (WHO) [2].

Brain tumor is one of the most important cancers causing death, represents the 17th most common cancer worldwide and accounts for 1%–2% of all tumors, this type of tumor has special importance, due to a significant increase in the incidence death rates from brain tumor in many developed countries. The peak rate of incidence of malignant brain tumors is seen in young children and in elderly individuals and the prognosis of this tumor is relatively poor, and for all ages, the average survival period is 9 months, and the 5-year survival rate is low [3].

The advancement in medical technologies helps the clinical experts to facilitate more efficient e-health care systems to the patients, Computer vision-based applications of biomedical imaging are gaining more importance as they provide recognition information to the radiologist for batter treatment-related problems [4]. Different medical imaging techniques and methods that include X-ray, Magnetic Resonance Imaging (MRIs), Ultrasound, and Computed Tomography (CT), have a great influence on the diagnosis and treatment process of patients [5][6].

MRI Scan
An MRI scanner can be used to take images of any part of the body (e.g., head, joints, abdomen, legs, etc.), in any imaging direction. MRI provides better soft tissue contrast than CT and can differentiate better between fat, water, muscle, and other soft tissue than CT (CT is usually better at imaging bones). These images provide information to physicians and can be useful in diagnosing a wide variety of diseases and conditions [7].

Computed tomography (CT) scan
CT scan uses x-rays to make detailed cross-sectional images of your brain and spinal cord CT scans are not used as often as MRI scans when looking at brain or spinal cord tumors, but they can be useful in some cases. They may be used if MRI is not an option, CT scans also show greater detail of the bone structures near the tumor [8].

When comparing the two technologies, it becomes clear that Some very small tumors sometimes do not show up on the CT scan, and this is why doctors almost always ordered MRI after a CT scan if there is concern for a brain tumor[9] , MRI scans are very good for looking at the brain and spinal cord and are considered the best way to look for tumors in these areas. The images they provide are usually more detailed than those from CT scans [10].
Deep learning is an innovative technique in machine learning. It is a powerful machine learning technology that has been applied as a solution to complex challenges that require high accuracy and sensitivity. The significant achievement was in medical devices, especially in the segmentation of the brain tumor. Thus, deep learning is a fast, fully automated, and competitive approach for brain tumor segmentation that guarantees accuracy and speed.

Deep learning is impeccable in analyzing images. In brain tumor segmentation, these techniques are high-ranking. Since one of the mandatory steps in processing images is semantic segmentation used in extracting the area of the brain that has been affected based on the MRIs [11].

The popularity of deep learning application in brain tumor analysis initially stemmed from workshops and conferences. Its potential applications were then published in journals. An enormous increase followed this in the number of research papers from 2015. Since then, the idea has become a significant topic of interest in subsequent journals and conferences [12]. Figure 2.1 shows in the development of deep learning applications to brain tumor analysis.

Figure 2.1. The application of deep learning in brain tumor analysis

Like any other new and complex technology, implementation of deep learning to analyze images in brain tumor evaluation has been accompanied by numerous challenges. Such drawbacks include the unavailability of large datasets used in training. Additionally, the whole process of labeling tumor images is fast, and the level of expertise needed to execute the task is high. Lastly, there is a risk of hiding vital information from another region as a result of the repetitive use of a fixed kernel size to slice the images. These challenges are summarized in figure 2.2 [13].

Figure 2.2. Challenges in the implementation of deep learning in tumor analysis

A Convolutional Neural Network or CNN is a deep learning neural network designed for processing a structured array of data such as images. They are widely used in computer vision and have become the state of the art for many visual applications such as image classification. Also, CNN is a unique machine learning structure originally modeled on the human visual cortex. Moreover, CNN is a powerful technology used for brain tumors detection. They simply involve analyzing features derived from the image to perform tasks such as classification of tumors.

The architecture of CNN is a multi-layered feed-forward neural network. It picks up patterns in the input image where an image is 784 pixels. The core entity of a neural network is where the processing takes place, each of the 784 pixels is fed to a neuron in the first layer of our neural network which forms the input layer. On the other end, we have the output layer, each neuron represents a digit with the hidden layers exiting between them. The information is transferred from one layer to another.
2.2 Related Work
In [14], MRI images are identified and categorized using an automated method. The automated method employs Super Pixel Technique and their individual classification. The classification of the Super Pixels involves a comparison of classifier of Extremely Randomized Tress (ERT) with SVM into normal and tumor. The stated technique has two datasets including BRATS 2012 dataset and 19 MRI FLIAR images. The results obtained show that the method has good performance using ERT classifier.

In [15], a CNN with 3*3 small kernels are used to identify tumor based on a method of automatic classification. In the method, a portion of enhancing region in dice similarity, core, coefficient matrix (0.88, 0.83, 0.77), and BRATS Challenge 2013 are obtained simultaneously.

In [16], Normal and MS tumors are diagnosed simultaneously using Alexnet model CNN. The results show the CNN classified 98.67% accurately into their respective classes.

In [17], a proposal for the use of multi-stage Fuzzy C-Means (FCM) in segmenting brain tumors from the MRI images was made.

In [18], the segmentation and classification were proposed to use an effective and efficient technique that used CNN. The technique extract features using Image-Net and gives 97.5% accurate results for the classification and 84% accurate results for the segmentation.

In [19], the results from multiphase MRI images in grading of tumors were studied and compared. The comparison focused on the results of base neural networks and deep learning structures. The findings of the comparison show that the performance of network based on their specificity and sensitivity of CNN improved by 18% as compared to neutral networks.
In [20], a technique known as deep learning-based supervised is used to detect changes in synthetic aperture radar (SAR) images. The stated technique provides a dataset with a right volume of data and variety for DBN training based on input images and resultant images from morphological operator applications on the images. The method has a detection performance that shows the suitability of algorithms based on deep learning for solving problems arising from change detections.

In [21], a proposal for a brain tumor that is automatic and complete is made based on DNN. The proposal advocate for networks that are designed to work in high glioblastoma and low-grade disease images. The paper also presents a new CNN architecture, which is cascading and uses the CNN output as an additional information source for the next CNN.

In [22], the Fuzzy C-Means (FCM) segmentation is applied to separate the tumor and non-tumor region of brain. Also, wavelet feature is extracted by using multilevel Discrete Wavelet Transform (DWT). An accuracy rate of 96.97% in the analysis of DNN based brain tumor classification but the complexity is very high, and performance is very poor.

In [23], new multi-fractal (MultiFD) feature extraction and improved AdaBoost classification schemes are used to detect and segment the brain tumor. The improved AdaBoost classification methods are used to find the given brain tissue is tumor or non-tumor tissue. Complexity is high but the accuracy is low.

In [24] a proposal for brain tumors classification is achieved by the convolutional neural network (CNN). The developed network is simpler than already-existing pre-trained networks, and it was tested on T1-weighted contrast-enhanced magnetic resonance images. network performance is estimated by four approaches: combinations of two 10-fold cross-validation methods and two databases.
. The best result is for 10-fold cross-validation was achieved for the record-wise method and, for the augmented dataset, and the accuracy was 96.56%.
In [25], SVM is a supervised classifier with associated learning algorithm based on the training samples, attempts to minimize the bound on the generalization error. The generalization error is the error made by the learning machine on the test data not used during training phase. Thus, the SVM always performs well when applied to data which is outside the training set. This approach significantly reduces the complexity and computation in solving the problem of classification. The advantages of this technique are high generalization performance, results are very accurate, and works well on high dimensional feature space. But it has some disadvantages such as long training time and highly depending on the data size.
In [26], Image classification is the primary domain, in which deep neural networks play the most important role of medical image analysis. The image classification accepts the given input images and produces output classification for identifying whether the disease is present or not, Also Instances can be classified by more than one output.

2.3 Proposed & Similar System Comparison
Your system System 1 System 2 System 3
Problem solved The biggest problem with classifying and segmenting the MRI images with some neural networks lies in the number of images in the database. This approach significantly reduces the complexity and computation in solving the problem of classification. In the area function approximation, it can be applied to make predictions and pattern recognition (it has been applied repeatedly in medicine for detecting cancer and other abnormalities in the human system).
Domain & Users radiologists in medical diagnostics. Brain images detection. This article presents a neural network approach for user modelling.
Design methods neural networks, used whole images as input for classification, and tested their networks with a k-fold designed SVM is a supervised classifier with associated learning algorithm based on the training samples Compares various results of deep leering structure and baseline Neural Network by
Software & hardware Algorithm requires many less resources for both training and implementing.
The algorithm can be on mobile platforms. Attempts to minimize the bound on the generalization error. Classified tumors with different grades based on SVM using features of 38 first-order or second-order statistic measurement.
Output images two times smaller than the provided input. Results are very accurate. Results at different layers show that the tumor feature can be closely resembled by the learned kernel
features The execution speed was quite good with an average than 15 MS per image.
Effective decision support tool for radiologists in medical diagnostics. high generalization performance, works well on high dimensional feature space. Used for classification or regression, able to present Boolean functions.
limitations The database smaller than database generally used in the field.
Using a simpler network requires fewer resources for training and implementation.
The network was trained and tested on a single graphical processing unit (GPU), CUDA device, GeForce. Long training time and highly depending on the data size Understating the structure of an algorithm is difficult.
Too many attributes can result in overfitting.
……..

3. System Analysis:
Analysis is the process of breaking a complex topic or substance into smaller parts to gain a better understanding of it. It may involve the examination and evaluation of the relevant information to select the best course of action from among various alternatives. In your analysis, you should explain the following: .

3.1 Requirements specification
You should Define the project goals in terms of functionalities and operations of the intended application. This may be done in the light of the end-user information needs. So, based on nature of the project, this section should include at least one of the following:
â–  Description of the methods that have been used to gather the project requirements & data such as: surveys, questionnaires, interviews, Investigation, Sampling, observations or any other methods.
â–  A structured Algorithm listing accompanied with a definite description of the scope and objectives of the problem under study. The listing should clearly define the set of inputs and the expected outputs.

CHAPTER 4 system design
4.1 system architecture
4.2

 
References
[1] Singh, L.; Chetty, G.; Sharma, D. A novel machine learning approach for detecting the brain abnormalities from MRI structural images. In IAPR International Conference on Pattern Recognition in Bioinformatics; Springer: Berlin, Germany, 2012; pp. 94– 105.

[2] Kleihues, P.; Burger, P.C.; Scheithauer, B.W. The new WHO classification of brain tumors. Brain Pathol. 1993, 3, 255–268[CrossRef].

[3] study to look at the histopathological pattern of these tumors at King Fahad Hospital (KFH) in Madinah region over 12 years. The basic demographic data were collected, and the tumors were studied under the guidelines of the WHO 2007 classification.

[4] Zhao, X.; Wu, Y.; Song, G.; Li, Z.; Zhang, Y.; Fan, Y. A deep learning model integrating FCNNs and CRFs for brain tumor segmentation. Med. Image Anal. 2018, 43, 98–111. [CrossRef] [PubMed].

[5] Singh, N.; Jindal, A. Ultra sonogram images for thyroid segmentation and texture classification in diagnosis of malignant (cancerous) or benign (non-cancerous) nodules. Int. J. Eng. Innov. Technol. 2012, 1, 202–206.

[6] Christ, M.C.J.; Sivagowri, S.; Babu, P.G. Segmentation of brain tumors using Meta heuristic algorithms. Open J. Commun. Soft. 2014, 1, 1–10. [CrossRef].

[7]https://www.fda.gov/radiation-emitting-products/mri-magneticresonance-imaging/benefits-and-risks.

[8]https://www.cancer.org/cancer/brain-spinal-cord-tumors-adults/detectiondiagnosis-staging/how-diagnosed.html.

[9] https://answers.zocdoc.com/details/.
[10] https://www.cancer.org/cancer/brain-spinal-cord-tumors-adults/detection- .diagnosis-staging/how-diagnosed.html.

[11] Sheller, M.J.; Reina, G.A.; Edwards, B.; Martin, J.; Bakas, S. Multi-institutional deep learning modeling without sharing patient data: A feasibility study on brain tumor segmentation. In International MICCAI Brainlesion Workshop; Springer: Berlin, Germany, 2018; pp. 92–104.

[12] Mittal, M.; Goyal, L.M.; Kaur, S.; Kaur, I.; Verma, A.; Hemanth, D.J. Deep learning based enhanced tumor segmentation approach for MR brain images. Appl. Soft Comput. 2019, 78, 346–354. [CrossRef].

[13] Çiçek,Ö.;Abdulkadir,A.;Lienkamp,S.S.;Brox,T.;Ronneberger,O.3DU-Net:Learningdensevolumetric segmentation from sparse annotation. In Proceedings of the International Conference on Medical Image Computing and Computer-Assisted Intervention, Athens, Greece, 17–21 October 2016; pp. 424–432.

[14] M. Soltaninejad, et al, Automated brain tumour detection and segmentation using superpixel-based extremely randomized trees in FLAIR MRI, International journal of computer assisted radiology and surgery, 12(2), pp. 183-203, 2017.

[15] S. Pereira, et al, Brain tumor segmentation using convolutional neural networks in MRI images, IEEE transactions on medical imaging, 35(5), pp. 1240-1251, 2016.

[16] Halimeh Siar, Mohammad Teshnehlab, Diagnosing and Classification Tumors and MS Simultaneous of Magnetic Resonance Images Using Convolution Neural Network, 7th Iranian Joint Congress on Fuzzy and Intelligent Systems (CFIS), 2019.

[17] L. Szilagyi, et al, Automatic brain tumor segmentation in multispectral
MRI volumes using a fuzzy c-means cascade algorithm, In 2015 12th
international conference on fuzzy systems and knowledge discovery
(FSKD), IEEE, pp. 285-291, 2015.

[18] Y. Xu, et al, Deep convolutional activation features for large-scale brain
tumor histopathology image classification and segmentation, In 2015
IEEE international conference on acoustics, speech and signal processing
(ICASSP), pp. 947-951, 2015.

[19] Y. Pan, et al, Brain tumor grading based on neural networks and convolutional
neural networks, In 2015 37th Annual International Conference
of the IEEE Engineering in Medicine and Biology Society (EMBC), pp.
699-702, 2015.

[20] F. Samadi, G. Akbarizadeh, et al, Change Detection in SAR Images using Deep Belief Network: a New Training Approach based on Morphological Images, IET Image Processing, 2019.

[21] M. Havaei, et al, Brain tumor segmentation with deep neural networks, Medical image analysis, 35, 18-31, 2017.

[9] Mohsen H et al. Classification using Deep Learning Neural Networks for Brain Tumors. Future Computing and Informatics. 2017:1-4.

[22] Islam A. et al. Multi-fractal Texture Estimation for Detection and Segmentation of Brain Tumors. IEEE.

[25] Parveen, Amritpal Singh, Detection of Brain Tumor in MRI Images, using Combination of Fuzzy C-Means and SVM, IEEE Conference on Signal Processing and Integrated Networks (SPIN), pg. no. 98-102, 2015.
[26] Y. Pan, et al, Brain tumor grading based on neural networks and convolutional
neural networks, In 2015 37th Annual International Conference
of the IEEE Engineering in Medicine and Biology Society (EMBC), pp.
699-702, 2015.

Appendix

Insert your appendix (appendices) here. The appendix should be used to provide any useful information, material, or derivations that is relevant to the project but should not be written in the report body in order not to interrupt the flow of the text.

The post AUTOMATED DIAGNOSIS FOR BRAIN TUMOR BASED ON MRI IMAGES PROJECT first appeared on COMPLIANT PAPERS.

GET HELP WITH YOUR PAPERS

GET THIS ANSWER FROM EXPERTS NOW

WhatsApp
Hello! Need help with your assignments? We are here
Don`t copy text!