Brain Tumor Segmentation and Classification- A Review Leena.B1, Dr.A.N.Jayanthi2
Department of ECE 1, 2
Assistant Professor, Associate Professor 1, 2
KGiSL Institute of Technology, Sri Ramakrishna Institute of Technology1, 2 Coimbatore, India.1, 2
Abstract
The current days Medical image processing is a quickly developing also a focusing area. Medical image techniques are used to detect and treatment of diseases. The medical field is growing at a rapid pace with new diseases are cropping up daily with need for invention of appropriate course of treatment. Medical imaging techniques are used to film the internal parts of the human body for clinical diagnosis in tumor diagnosis. The tumor part in the radiological image format like X-ray, MRI and CT scans are extracted using segmentation and various classification techniques. Various articles based on the MRI and CT images are segmentation and various classification techniques discussed here. To analyze the excellent of the image, image de-noising procedure followed by picture segmentation and image classification methods are implemented. Thus, in this survey, different approaches of classification and segmentation of tumour image is scrutinized.
Keywords: Computed Tomography, Magnetic Resonance Imaging, Brain Tumor, tumor tissue image segmentation, image classification.
I. INTRODUCTION
Image processing is an effective area of research and the medical image processing is a interesting field in image processing where the images of different quality of the inner portions of human body are processed to discover any complications, observe that the biomedical images are extensively used for many purposes in these existing medical cases.
Brain is the soft tissue, which is held within the protective shield known as skull. Brain cancer is an unusual development in the human brain, which may cause serious problems in the human body and subjected the tumor affected persons to life threatening. The automated notice of brain tumor is mainly enable the division process. The Section of brain images used to extract the abnormal portion of the image that is the tumor. From the extracted size and shape of the abnormal portion, it is effective to determine the nature of the treatment that can be given to the patients. Numerous research articles have been reported in the literature on different approaches to image division and image classification. This review analysis provides a comprehensive overview of the current methods for brain division and tumor classification. Analysing images in the computer will help radiologists to detect the important or suspicious region in the image, so as to diagnose diseases.
Table 1 Comparison of CT and MRI in brain Tumor Assessment
Modality Indications Limitations
CT Quick picture time
Small amount of scanning
Greater spatial resolution
extra-axis brain tumor evaluation
Elevated in calculations,
skull erosion,
Acquisition of only one aircraft and most of the time the risk of nonisotropic X-ray radiation
Poor tissue
representation
infiltration, destruction
MRI Greater noticing of mass effects and atrophy
Definition of Higher Neurophysiology (Tissue Difference)
Accurate detection of tumor vascularity (in various aircraft acquisitions).
Precise selection of edema and compression effects
Bad diagnosis of calcification and bone itching
Impossible in introspective assessment
Low spatial reliability
Some time sequences are very time consuming
In medical science, these images play a vital role. The use case like MRI- Magnetic Resonance Imaging captures the information of the inside formation of human brain Also used to depict other parts of the body.MRI data are manually analyzed by medical professionals in brain tumors, which is a difficult and complicated process.
Here, mainly review the last systems, identify the issues in it and assist it in move the moreover research successfully to design an effective proceed towards for tumor classification and segmentation using MRI with novel machine learning schemes. One such basic and life-imperiling disease is brain tumor which is an unusual Brain tissues growth within the brain. Association of brain tumour an exploration because of the difficulties in configuration of the brain. Existing technique of MRI brain tumor noticing, segmentation is reviewed in the review paper along with its benefits and drawbacks. Although there are numerous techniques available to diagnose a brain tumor, early detection of a brain tumor is important to save the patient's life.MRI diagnosis of a brain tumor provides an accurate view of the brain tumor. The division and classification of the brain plays an important role in the diagnosis. Proper sectioning of the brain tumor can help with surgery or removal of the brain tumor. The different methods of brain tumor classification and diagnosis are as follows, which provide detailed information on brain tumor diagnosis, segmentation and classification. Table 1 illustrates the difference between MRI and CT image processing for brain cancer segmentation and classification.
MRI Image segmentation and classification is used to detect cancerous cells from medical images. Analysing medical images for the purpose of Computer-Aided Diagnosis (CAD) and planning treatment, makes segmentation a preliminary stage for visualisation or quantification. Presently, for medical CT and MRI images, several methods are employed for segmentation. This chapter deals with literature review of image preprocessing techniques and segmentation algorithms for MRI brain tumor images, which are closely linked to the topic of this thesis. Several research works are being carried out by various researchers to find the tumor affected region in MRI brain images. It provides different analyses for disease prediction algorithms, and the challenges and problems with medical images. A wide range of suggestions and future studies for some problems and published articles by different researchers are analysed and presented in this chapter.
The reminder of this paper is arrange as proceed division 2, brain cancer segmentation and categorization its related work, division 3 discussed to research gap of conventional
approach. Finally, section four provides the concluding remarks And the future purpose of the work.
II. LITERATURE REVIEW
Menze et al (2014) has proposed to a structure as automated brain tumor separation from MR images. Because ability to the expansion about edema crucial as analysis, plan,alsocure edema can be diagnosed simultaneously with tumor segregation, while multiple tumor segregation methods enhance the severity formed through cadolinium contradiction agent of T1. Weighted picture no more require advanced picture channels different from method used here. Alone input needed as separation process is the T2 MR picture channel, although this used for additional upgraded picture as advanced tissue separation. This section structure has three levels. Initially, detection of untypical section is done used to recorded brain picture exemplary as an active brain.
Strong opinion through point also scattering about clusters of different tissue on brain intensities are then using into regulate extreme characteristics of various tissue cases. this second point, they determine taken awayT2 picture intensities whether edema appears in sync for the cancer irregular areas. Lastly, then Apply spatial barriers toward diagnosed tumor areas.
Szilagyi et al. (2015) have Traditional fuzzy C-means algorithms (FCM) expressed the robust separation procedure established the increase of clustering procedure. In homogeneities imperfection associated with the acquisition sequences provides major contribution to MRI brain image data. discussed FCM clustering method more deeply in their research work. Partition matrix and random initialisation are some of the disadvantages of FCM which makes clustering results more inconsistent. An alternative method called Subtractive Clustering (SC) was considered, however, the numbers of clusters were unknown. The authors proposed a new hybrid method which combines FCM and SC called as Subtractive Fuzzy C Means (SFCM) to overcome the disadvantage of FCM and SC. The experimental results showed that SFCM method results produces much better clustering results than FCM algorithm indices. neighbor gravity that depends on connective point and featured about the closing picture element using toward dramatically develop the separation act.
Selvathi.D also Henry Selvaraj (2018) have proposed Framework for vague information integration toward automatically separate cancer region about the MRI image, such as T1-weighted, T2-weighted, and proton-density (PD) images. Several methods exist for preprocessing of images and all these methods have their own advantages and disadvantages. Contrast enhancement, boundary detection and removal of unwanted noisy pixels are some of the aspects in preprocessing. Without losing any information, noises in images are removed for further detection or identification of diseases or affected region.Previous expertise of cancers defined through radiologists for various category about MRI can be useful in guiding an automated and accurate segmentation.
However, the terms used by radiologists vary over the duration of the image signal.
As a result of these interpretations, a type of MRI sequence was constructed to modify the vague specimens. They also gave a brief introduction about MRI characteristics and its noises. Noises corrupt the images during image acquisition process, which degrades the image quality. Most of the denoising methods are dealt with Rician noise with spatial
uniform noise distribution in MRI images.This section lastly situated about combination on various obscure instruction earn taken away various categorize picture in MRI.
Khotanlou Hassan, et al (2009) has submitted a new general method as classifying 3D MRI from brain tumor., which suitable toward various categorizing of tumors. Early, brain was dissecting used a strong modern toward existence of cancer. The early cancer diagnosis was made situated to the selection of unbalanced regions in relation toward near similarity of brain position also ambiguous analysis. Output creates initiation to the segregation system situated to a sequence about distorted model relationships, which leads toward the correct separation about cancers, also precision and volatility were take to account at all levels, using appropriate ambiguous models.
Gui Laura et al (2012) proposed a technique bellow morphology based and presented how to extract easily information and various features from the images segmentation for detection of brain tumor. The brain segment (MBRASE), presents MRI image for tumor detection using morphological filtering which is removed from the input image and also presented based on morphological operation and segmentation algorithm. The experiment was conducted for the detection of tumor by using morphological operation to image and later segmentation in particular watershed for s canned MRI image from human brain for different samples the algorithm.
Iqbal Sajid et al. (2018) have proposed the structure of the clinical picture analysis system to follow brain tumor division and presented two methods that are used to plan the image in MRI. By this assist of design how to distinguish the boundaries of brain cancer and calculate the genuine area of tumor and also the f-transform is worn to give the definite information like rebuilt of misplaced edges and extracting the hushed edges. Accuracy and simplicity in an MRI Images is reliant on each other.
Table 2 Earlier procedure and techniques for segmentation of brain tumor
Author Methods Inference & Remarks
Menze et al.2014 A structure for automatic brain tumor separation from MR images.
1. Tumor segmentation and Diagnosis of edema is done at same time.
2. Inefficient to minimize Computational time
Szilagyi et al. 2015 A robust segmentation technique
1. Improves the performance of segmentation.
2. Critical task to improve image quality.
Selvathi.D and Henry
Selvaraj (2018) A structure of ambiguous information merging
.
1.Automatic tumor
segmentation from
multispectral MRI.
2. Lacking of accuracy.
Khotanlou Hassan et al (2009)
A appropriate Fuzzy Model. 1. Segmenting different types of brain tumors in 3D MRI.
2. Hurdle in improving In homogeneity and reducing blur images.
Gui Laura et al (2012) MBRASE 1. Automatic seperation of
T1 weighted MR image data from brain tumor.
2. Hurdle in improving quality of images.
Iqbal Sajid et al (2018) Structure of the clinical image analysis system.
1. Segmentation of tumors.
2. Obstacle to eliminate noise.
Sompong Chaiyanan and Sartra Wongthanavasu (2017)
Seed tumor separation method based on a cellular automata (CA).
1. Areas with minimal user contact are solid tumors.
2. Difficult with reduce computational complexity.
Sompong Chaiyanan and Sartra Wongthanavasu (2017) have proposed a quick and robust practical tool for minimally invasive solid tumor dissection to assist physicians and researchers in evaluating the response to radiosurgery planning and treatment. In particular, a cellular automata (CA) based seed cancer separation technique for differentiated T1 weighted MRI pictures, it systematize return rate (VOI) also seed picking. First off to start CA-based connection on segment with chart theoretical process.
In this regard, they chief of state transformation the objective of CA to workout correct short line result. moreover, a sensation restriction has been imported to suit a multivariate cancer separation issue, also indirect surface area formed in the cancer expectation plan- formulate against CA case into appoint structural softness. Satisfactory report into begin the procedure collected boundary drawn from buyer to peak diameter of the tumor, on accordance for medical habit. Moreover, a method depend on CA provided toward improve harmful also cancer tissue contentment, whatever is important as comprehensive evaluation about the emission healing feedback. The table 2 discussed about the researchers’ survey of different types of segmentation algorithms and preprocessing techniques and given a short summary of their advantages and disadvantages.
watani Jun et al (2015) have proposed the according to medical practice, enough information is collected from the user to start the process through line tired with biggest diameter about cancer. Furthermore, a method placed about CA provided to improve harmful also cancer tissue contentment, whatever is important as comprehensive evaluation about the emission healing feedback. Wenlu et al. (2015) have proposed that the Neurological networks are generally noted in the investigation association about clinical picturing, which focuses on the recent neurological network advances in computer diagnostics, clinical imaging, and marginal find into study of visual content,
also clinical picture recording as front-processing also back-processing, raising awareness about how neural networks can be enforced to these field and in case a base for further investigation and practicable improvement.
A hybrid method for separating WM into brain images using a granular rough set and blurred threshold provided by Senthilkumaran N and Rajesh. it takes the inertia factor into the algorithm, thus, optimal threshold was acquired for image segmentation. This algorithm had more advantages such as high segmentation speed, stability and converges easily to get the optimal solution. The granular approximate synthesis approach for brain image segmentation combines the obscure threshold for brain WM segmentation and the results show performance.
Most of the methods proposed in the literature have the disadvantage of losing the background image. A hybrid model was proposed by Sharif Muhammad et al. (2018) which identify the ROI Using the combined results of the threshold section and image functions.
Initially, an abnormal brain MR image is processed with morphological functions such as ossification-based segmentation and erosion. Furthermore, in order to protect the background and correctly identify the tumor area, the separated result images are both combined with the original MR image.
Benson.C.C et al 2015 have presented A new approach called hybrid division inspired by mathematical morphological operators and morphological aquatic division. This approach benefits from the complementarity between these two approaches. Image drivers almost exclude tumor area and may eventually affect healthy structures, while the hydrocephalus system provides details of different brain structures, so the combination of these two approaches significantly improves the separation of the tumor and tumor zone.
A fresh hybrid method based on the SVM and fuzzy c-means as brain cancer categorization is submitted by Srinivas and Sasibhushana Rao (2019). Introduced the evaluation set of rules with a variety of investigative and diagnostic scenarios in the directions of large-scale retrieval, which can more improve the enactment of medical image investigation. SVM and fuzzy c-means method mix of the problem solving, a hybrid procedure about predicting tumors in brain. Fuzzy C-Means (FCM) flock using as segment to detect suspicious area in brain MRI image. The SVM method used toward brain MRI pictures are segregate that provides an exact also highly valid decision as classifying the MRI images from brain.
Table 3 Conventional Hybrid methods and techniques for brain tumors segmentation
Author Methods Inference & Remarks
watani Jun et al. (2015) Single WM voxel 1. The global least is generated in WM after the watershed algorithm is applied.
2. Inefficient to minimize computational time.
Wenlu et al. (2015) Artificial Neural Network used to analysis medical image
1. computer-assist analysis, medical picture separation also edge detection and medical image booking field to focus.
2. Lacking of accuracy and efficiency.
Senthilkumaran.N and Rajesh.R (2011)
A hybrid method for WM separation
1. Brain image segmentation and brain WM separation.
2. Difficult to detect edge errors.
Sharif Muhammad et al.
(2018)
Threshold segmentation and Morphological operations.
1. Identifies the ROI using fused results.
2. Unable to improve image quality and reduce blur images.
Benson et al 2015 Morphological watershed 1. Both segmentation and
segmentation. extraction of the tumor zone have been improved.
2. Critical task to improve accuracy & homogeneity.
Srinivas and Sasibhushana Rao (2019)
SVM and FCM to use the new compound procedure.
1. Segmentation, classification and prediction of brain tumor.
2.Fails to reduce
computational complexity.
Sharma Manorama et al. (2018)
Framework of the segmentation and an automatic detection of brain tumor.
1. Observation and segmentation of brain tumors.
2. Obstacle to improve image quality.
The table 3 explain the Conventional Hybrid methods also separation techniques about cancer in brain. Manual Diagnosis also segmentation of Brain MRI in Today's Brain MRI With numerous MRI scans per patient is difficult and subject to intermediate and internal observer diagnosis and segmentation variation. To overcome this, Sharma Manorama et al. (2018) approached automated brain tumor detection and division structure includes techniques ranging from skull removal to diagnosis and brain tumor dissection. Through pre- processing, image fusion and initial tumor strip classification, tumor segmentation based on the final hybrid intelligent fuzzy Hopfield neural network algorithm, and tumor area detection and extraction are achieved. Role Classifier an important one in implementation of Automatic System.This system using toward notice the brain cancer tissue taken away MRI brain pictures. Many columnist proposed different procedure of classification (Yin Zhong and Jianhua Zhang 2014, Lee Chi-Hoon et al 2014; Raju.A et al 2018; Sivaramakrishnan A et al (2014); Prastawa Marcel et al (2004).
Yin Zhong and Jianhua Zhang 2014 designed a classification method using Recursive Support vector machine (SVM RFE) based feature removal for genetic selection and taxonomy,They are integrated into a standard structure.SyedAqhsa Q and K. Narayanan (2014) drawed Artificial Neural Networks (ANN)-Many layer perceptron neural network for classify the image features.
Lee Chi-Hoon et al (2014) represented Support Vector Machine used for preparing many Attempts were made to locate the multi-aspect vectors and separating hyper plane based on n-peptide compounds with great distance.Wu & Moon (2012) classified a glioma tumors using Brain electrical activity mapping. Huang Desheng et al.(2009) Linear Discrimination Analysis (LDA) for the classification of linear, nonlinear, low- and high- grade tumors, specific techniques on the lower square support vector machine (LS-SVM).
Prastawa Marcel et al (2004) Voxel and Geometric Model classifying are noticed Table 4 Conventional methods and techniques for brain tumors classification
Author Methods Inference & Remarks
Yin Zhong and Jianhua Zhang (2014)
SVM RFE and GA using Recursive feature elimination.
Toward resolves the above mentioned excellent uplifting planes, it clarifies the cumulative duplicate code issue. Using toward obtain excellent values. Its Reach huge analysis efficiencies along genes.
The completion do perfect satisfactory.
Lee Chi-Hoon et al (2014)
Support Vector
Machines(SVM)
To predict sub cellular localization.
Chato Lina and Shahram Latifi (2017)
(MLFFNN)Multi-Layer Feed Forward Neural Network, and SVM
Utilize toward arrange tumor areas taken away none-tumor areas.
(Raju.A et al (2018)) Bayesian Model Individual area committed Models class mostly confer towards the put away from sample classes studied.
Sivaramakrishnan, A et al (2014).
Maximization
Method of 3D-Assumption, Hidden Markov Model
To categorize the tumor areas.
Prastawa Marcel et al (2004)
Voxel Approach, Geometric Model.
To categorize the tumor areas from non-tumor areas.
Syed Aqhsa Q and K.
Narayanan (2014)
Artificial Neural Networks (ANN) - Multilayer preceptron neural network (MPNN).
Towards arrange the cancer featured exact on the spectra.
Prastawa Marcel et al (2004)
Supervised voxal
Classification.
The tumors classified successfully.
Alfonse M & Salem A. B. M (2016)
Support Vector Machines (SVMs), Decision Tree(DT)
90.8% less taken away huge quality cancer also 85.6%
low from huge invasive cancer also clearly restricted. A SVM capability was verified toward provide great act even in finite teaching models.
Morais C. L. et al (2019)
Quadratic Discriminate Analysis and Support Vector Machine
A SVM allocation is linked to the QDA-based classification for better tumor profile.
Huang Desheng et al.(2009), Çınarer G &
Emiroğlu B. G (2019)
Linear Discriminant Analysis, Least Squares Support Vector Machines
The classifiers randomly divided the dataset into 100 tiers of training and test sets
(LS-SVM), Linear Kernel Techniques
Wu & Moon (2012) Multi-Scale -Based Classification.
It using in to naturally selected subgroups about areas created by Watershed segmentation system.
Iqbal Sajid et al (2018)
Statistical Classification method, magnetic resonance imaging features to use Computer-assisted brain tumor type discrimination
Categorize the tumor is benign, Malignant or usual.
Menze et al (2007) Atlas-moderated automatic tissue classification technique, Expectation Maximization Segmentation (EMS).
To segment tumor and edema successfully.
Sharma R et al (2019) (BPNN), and Ant Colony Optimization technique (ACO)
Masses were extracted completely.
Sharma R et al (2019) Rough Set Particle Swarm Optimization
It can create the most common Results rules and new excellent classification quality Samples.
The table 4 discussed about the researchers’ survey of different types of classification techniques and given a short summary of their advantages and disadvantages. That analysis different automated identifying procedure of brain cancer by MRI have on prepared also related as higher than two decades .It using towards target on the future of advances in deal with medical picture in medication and medical management. The proposed system express a lot procedure deal with clinical pictures also consider the demand also characteristics about procedure on brain cancer identification .That study using toward provide additional knowledge on brain cancer diagnosis also seperation. This aim in the analysis of total technological related toward brain tumors, taken away MRI to deal with clinical picture
RESEARCH GAP
This analysis provided the review on the details associated with picture segmentation and categorization techniques, their challenges in addition of benefits also shortcomings. It also addressed the detection of tumor employing a variety of algorithms. This way, the segmentation and classification employing various algorithms has been studied. From the review it has been found that several methods were used and several hypotheses were arrived in literature. Few methods require radiologist intervention while few are automatic. But the fully automatic brain cancer segmentation methodology take part in bigger position and usage of detached treatments. The conventional techniques were huge and out dated due to hardware implementation. Fitting to prove our objective also importance to suggest brain tumour implementation, knowledge of various researchers in the field is examined and reviewed. The techniques make use of the segmentation of the brain cancer and detection edges played main part of tumor analysis in brain. Existing algorithms and techniques used identification and the tumors segmentation in brain of the medical picture processing have drawbacks such as large time consuming, and less efficiency. Various technique used for detection of brain tumor, The division and classification with advantages and disadvantages in Table 2, Table 3 and Table 4 are explored above. So various procedure as brain cancer
detection, separation also allocation have been described above, but it has the following drawbacks.
It has low knowledge of splitting diseased tumor tissue from normal tissue.
Abnormalities classification is not predictable
Low level of accuracy in segmentation
This method requires evaluation and in most cases does not give accurate results.
It takes more time
III. CONCLUSION
In the biomedical diagnosis, image Processing is the most dominant field. The system handle many dimensional, complex also vast details. Pre-processing, feature extraction and classification design should be optimal. Advance technology need to make the design as required and the computing perform effectively. The implementation of algorithms is increasing day by day due to several issues involved. New methods, technologies and techniques are innovated to optimize the process. This research work aims to find the pre- processing, segmentation and classification algorithm issues in MRI image.
References
[1] Menze, Bjoern H., et al. "The multimodal brain tumor image segmentation benchmark (BRATS)." IEEE transactions on medical imaging 34.10 (2014): 1993-2024.
[2] Szilagyi, L., Lefkovits, L., & Benyo, B. (2015, August). 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) (pp. 285-291). IEEE.
[3] Selvathi, D., and Henry Selvaraj. "Segmentation Of Brain Tumor Tissues In Mr Images Using Multiresolution Transforms And Random Forest Classifier With Adaboost Technique." 2018 26th International Conference on Systems Engineering (ICSEng).
IEEE, 2018.
[4] Khotanlou, H., Colliot, O., Atif, J., & Bloch, I. (2009). 3D brain tumor segmentation in MRI using fuzzy classification, symmetry analysis and spatially constrained deformable models. Fuzzy sets and systems, 160(10), 1457-1473.
[5] Gui, L., Lisowski, R., Faundez, T., Hüppi, P. S., Lazeyras, F., & Kocher, M. (2012).
Morphology-driven automatic segmentation of MR images of the neonatal brain. Medical image analysis, 16(8), 1565-1579.
[6] Iqbal, S., Ghani, M. U., Saba, T., & Rehman, A. (2018). Brain tumor segmentation in multi‐spectral MRI using convolutional neural networks (CNN). Microscopy research and technique, 81(4), 419-427.
[7] Sompong, C., & Wongthanavasu, S. (2017). An efficient brain tumor segmentation based on cellular automata and improved tumor-cut algorithm. Expert Systems with Applications, 72, 231-244.
[8] watani, J., Ishida, T., Donishi, T., Ukai, S., Shinosaki, K., Terada, M., & Kaneoke, Y.
(2015). Use of T1‐weighted/T2‐weighted magnetic resonance ratio images to elucidate changes in the schizophrenic brain. Brain and behavior, 5(10), e00399.
[9] Zhang, W., Li, R., Deng, H., Wang, L., Lin, W., Ji, S., & Shen, D. (2015). Deep convolutional neural networks for multi-modality isointense infant brain image segmentation. NeuroImage, 108, 214-224.
[10] Senthilkumaran, N., & Rajesh, R. (2011). Brain image segmentation. International journal of wisdom based computing, 1(3), 14-18.
[11] Sharif, M., Tanvir, U., Munir, E. U., Khan, M. A., & Yasmin, M. (2018). Brain tumor segmentation and classification by improved binomial thresholding and multi-features selection. Journal of Ambient Intelligence and Humanized Computing, 1-20.
[12] Benson, C. C., Lajish, V. L., & Rajamani, K. (2015, August). Brain tumor extraction from MRI brain images using marker based watershed algorithm. In 2015 International Conference on Advances in Computing, Communications and Informatics (ICACCI) (pp. 318-323). IEEE.
[13] Srinivas, B., & Rao, G. S. (2019). Performance evaluation of fuzzy C means segmentation and support vector machine classification for MRI brain tumor. In Soft computing for problem solving (pp. 355-367). Springer, Singapore.
[14] Sharma, M., Purohit, G. N., & Mukherjee, S. (2018). Information retrieves from brain MRI images for tumor detection using hybrid technique K-means and artificial neural network (KMANN). In Networking communication and data knowledge engineering (pp. 145-157). Springer, Singapore.
[15] Yin, Zhong, and Jianhua Zhang. "Operator functional state classification using least- square support vector machine based recursive feature elimination technique."
Computer methods and programs in biomedicine 113.1 (2014): 101-115.
[16] Syed, Aqhsa Q., and K. Narayanan. "Detection of tumor in MRI images using artificial neural networks." International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering 3.9 (2014): 11749-11754.
[17] Lee, C. H., Schmidt, M., Murtha, A., Bistritz, A., Sander, J., & Greiner, R. (2005, October). Segmenting brain tumors with conditional random fields and support vector machines. In International Workshop on Computer Vision for Biomedical Image Applications (pp. 469-478). Springer, Berlin, Heidelberg.
[18] Wu, W. J., Lin, S. W., & Moon, W. K. (2012). Combining support vector machine with genetic algorithm to classify ultrasound breast tumor images. Computerized Medical Imaging and Graphics, 36(8), 627-633.
[19] Huang, D., Quan, Y., He, M., & Zhou, B. (2009). Comparison of linear discriminant analysis methods for the classification of cancer based on gene expression data. Journal of experimental & clinical cancer research, 28(1), 149.
[20] Prastawa, M., Bullitt, E., Ho, S., & Gerig, G. (2004). A brain tumor segmentation framework based on outlier detection. Medical image analysis, 8(3), 275-283.
[21] Sapra, Pankaj, Rupinderpal Singh, and Shivani Khurana. "Brain tumor detection using neural network." International Journal of Science and Modern Engineering (IJISME) ISSN (2013): 2319-6386.
[22] Chato, Lina, and Shahram Latifi. "Machine learning and deep learning techniques to predict overall survival of brain tumor patients using MRI images." 2017 IEEE 17th International Conference on Bioinformatics and Bioengineering (BIBE). IEEE, 2017.
[23] Raju, A. R., Suresh, P., & Rao, R. R. (2018). Bayesian HCS-based multi-SVNN: A classification approach for brain tumor segmentation and classification using Bayesian fuzzy clustering. Biocybernetics and Biomedical Engineering, 38(3), 646-660.
[24] Sivaramakrishnan, A., M. Karnan, and R. Sivakumar. "Medical Image Analysis–A Review." Sivaramakrishnan et al,/(IJCSIT) International Journal of Computer Science and Information Technologies 5.1 (2014): 236-246.
[25] Kochar, Priya. "A Survey on Brain Tumor Detection and Classification System based on Artificial Neural Network." International Journal of Computer Applications 90.18 (2014).
[26] Alfonse, Marco, and Abdel-Badeeh M. Salem. "An automatic classification of brain tumors through MRI using support vector machine." Egy. Comp. Sci. J 40.3 (2016).
[27] Morais, C. L., Lima, K. M., & Martin, F. L. (2019). Uncertainty estimation and misclassification probability for classification models based on discriminant analysis and support vector machines. Analytica chimica acta, 1063, 40-46.
[28] Çınarer, G., & Emiroğlu, B. G. (2019, October). Classificatin of Brain Tumors by Machine Learning Algorithms. In 2019 3rd International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT) (pp. 1-4). IEEE.
[29] Iqbal, S., Khan, M. U. G., Saba, T., & Rehman, A. (2018). Computer-assisted brain tumor type discrimination using magnetic resonance imaging features. Biomedical Engineering Letters, 8(1), 5-28.
[30] Sharma, R., Marikkannu, P., & Sungheetha, A. (2019). Three-dimensional MRI brain tumour classification using hybrid ant colony optimisation and grey wolf optimiser with proximal support vector machine. International Journal of Biomedical Engineering and Technology, 29(1), 34-45.
[31] Sharma, R., Marikkannu, P., & Sungheetha, A. (2019). Three-dimensional MRI brain tumour classification using hybrid ant colony optimisation and grey wolf optimiser with proximal support vector machine. International Journal of Biomedical Engineering and Technology, 29(1), 34-45.
[32] Zhang, W., Li, R., Deng, H., Wang, L., Lin, W., Ji, S., & Shen, D. (2015). Deep convolutional neural networks for multi-modality isointense infant brain image segmentation. NeuroImage, 108, 214-224.
[33] Raju, A. R., Suresh, P., & Rao, R. R. (2018). Bayesian HCS-based multi-SVNN: A classification approach for brain tumor segmentation and classification using Bayesian fuzzy clustering. Biocybernetics and Biomedical Engineering, 38(3), 646-660.
