Brain Tumor Classification using CNN

Abstract

In this paper, we propose a brain tumor segmentation and classi?cation method for multi-modality magnetic resonance image scans. The data from multi-modal brain tumor segmentation challenge are utilized which are co-registered and skull stripped, and the histogram matching is performed with a reference volume of high contrast. We are detecting tumor by using preprocessing, segmentation, feature extraction, optimization and lastly classi?cation after that preprocessed image use to classify the tissue. We performed a leave-one out cross-validation and achieved 88 Dice overlap for the complete tumor region, 75 for the core tumor region and 95 for enhancing tumor region, which is higher than the Dice overlap reported.

Introduction

I. INTRODUCTION

The discovery and conclusion of cerebrum tumor from MRI is pivotal to diminish the pace of setbacks. Cerebrum tumor is dif?cult to fix, in light of the fact that the mind has a complex structure and the tissues are interconnected with each other in a complicated manner. In spite of many existing methodologies, vigorous and ef?cient division of cerebrum tumor is as yet a significant and testing task. Tumor division and classi?cation is a challenging task, because tumors vary in shape, appearance and location. It is hard to completely section and group cerebrum tumor from mono-methodology checks, on account of its confounded structure. X-ray gives the capacity to catch different pictures known as multimodality pictures, which can give the itemized structure of cerebrum to ef?ciently group the mind tumor. shows diverse MRI modalities of cerebrum.

II. PROBLEM STATEMENT

Image processing and deep learning techniques are used to segment brain MRI data.

A. Motivation

One of the most difficult tasks in medical image analysis is tumour segmentation the most prominent research in the field of medical image analysis focuses on brain tumors. detection of a Brain Tumor Early reducing the amount of pressure placed on human judgments. rapid communication allows for health treatment to be extended to far-flung locations early discovery reduces death rates.

III. LITERATURE SURVEY

Lina Chato , Erik Chow,” Wavelet Transform to Improve Accuracy of a Prediction Model for Overall Survival Time of Brain Tumor Patients Based On MRI Images Lina ”[1]

In this poster, denoising wavelet transform (DWT) method is proposed to improve the accuracy of a prediction model for overall survival time of brain tumor patients using Magnetic resonance imaging (MRI) images based on classi?cation approach. The BraTS dataset is used in this work. The histogram features are extracted from MRI images to train a prediction model using machine learning methods. As the dataset consists of only 163 samples, various machine learning methods have been used to develop an accurate prediction model. In general the MRI imaging system corrupted the MRI information with noise. The results show that the two dimension denoising wavelet transform method slightly improved the accuracy of a prediction model based on histogram features. The best accuracy is achieved by daubechies 4 level 4 (db4-L4) with a10 folds cross validation linear support vector Machine (SVM) when including patients’ age information. However, daubechies 2 level 1 and 3 (db2-L1, db2L3) with a 10 folds cross validation simple tree produce animprovedaccuracywhenthepatients’agedoesnotcombinedwithhistogramfeatures vector. When a 10 hold out validation method is used, the daubechies 2 level 3 (db2-L3)with simple tree achieves66.7 accuracy.

Linmin Pei, Syed M. S. Reza and Khan M. Iftekharuddin .” Improved Brain Tumor Growth Prediction and Segmentation in Longitud in al Brain MRI”,[2] In this work, we propose a novel method to improve the predication of brain tumor growth by fusing with the state-of-art tumor segmentation.

The Glioma Image Segmentation and Registration (GLISTR) is known for joint segmentation and deformable registration of brain scans as well as tumor growth prediction using MRI. This paper, for the ?rst time in literature, aims to improve the tumor growth prediction by integrating the growth patterns of different tissues such as necrosis, edema, and tumor obtained from GLISTR with our stochastic texture-based tumor segmentation methods using a joint label fusion (JLF) technique. We evaluate the proposed method using several adult longitudinal cases from the 2015 BRATS [1) dataset. The experimental results show difference of these tissues growth prediction by applying GLISTR and joint label fusion. ANOVA analysis suggests statistically improvement in the longitudinal tumor core prediction results.

G.Hemanth , M.Janardhan ,L.Sujihelen,” DESIGN AND IMPLEMENTING BRAIN TUMOR DETECTION USING MACHINE LEARNING APPROACH” [3] Nowadays, brain tumor detection has turner dupes’ general causality in the realm of health care. Brain tumor can be denoted as a malformed mass of tissue wherein the cells multiply abruptly and ceaselessly, that is there is no control over the growth of the cells. The process of Image segmentation is adopted for extracting abnormal tumor region within the brain. In the MRI (magnetic resonance image), segmentation of brain tissue holds very signi?cant in order to identify the presence of out lines concerning the brain tumor. There is a abundance of hidden information in stored in the Health care sector. With appropriate use of accurate data mining classi?cation techniques, early prediction of any disease can be effectively performed. In the medical ?eld, the techniques of ML (machine learning) and Data mining holds a signi?cant stand. Majority of which is adopted effectively. The research examines list of risk factors that are being traced out in brain tumor surveillance systems. Also the method proposed assures to be highly ef?cient and precise for brain tumor detection, classi?cation and segmentation. To achieve this precise automatic or semiautomatic methods are needed. The research proposes an automatic segmentation method that relies upon CNN (Convolution Neural Networks), determining small 3 x 3 kernels. By incorporating this single technique, segmentation and classi?cation is accomplished. CNN (a ML technique) from NN (Neural Networks)wherein it has layer based for results classi?cation. Various levels involved in the proposed mechanisms are: 1. Data collection, 2. Pre-processing, 3. Average ?ltering, 4. segmentation, 5. feature extraction, 6. CNN via classi?cation and identi?cation. By utilizing the DM (data mining) techniques, signi?cant relations and patterns from the data can be extracted. The techniques of ML (machine learning) and Data mining are being effectively employed for brain tumor detection and prevention at an early stage.

Parveen , Amritpal singh,” Detection of Brain Tumor in MRI Images, using Combination of Fuzzy C-Means and SVM”[4], MRI is the most important technique, in detecting the brain tumor. In this paper data mining methods are used for classi?cation of MRI images. A new hybrid technique based on the support vector machine(SVM)and fuzzy c-means for brain tumor classi?cation is proposed. The purposed algorithm is a combination of support vector machine (SVM) and fuzzy cmeans, a hybrid technique for prediction of brain tumor. In this algorithm the image is enhanced using enhancement techniques such as contrast improvement, and mid-range stretch. Double thresholding and morphological operations are used for skull striping. Fuzzy c-means (FCM) clustering is used for the segmentation of the image to detect the suspicious region in brain MRI image. Grey level run length matrix (GLRLM) is used for extraction of feature from the brain image, after which SVM technique is applied to classify the brain MRI images, which provide accurate and more effective result for classi?cation of brain MRI images.

IV. SYSTEM ARCHITECTURE

A. Module

Pre-processing

Feature Extraction

Classifification

B. Algorithm

1. CNN: A convolution (CNN/ConvNet) is a form of deep learning model used to evaluate visual imagery in machine learning. When we conceive about neural networks, we usually think of multiplications, but this isn't the case with Convolution. It employs a technique known as Convolution. Pooling layer is a mathematical function on two functions that yields a third function that explains how the shape of one is changed by the other. Multiple layers of artificial neurons make up convolution neural network. Connections between neurons are mathematical in nature that calculate the weight value of various inputs and output an activity value, similar to their natural counterparts. Each layer creates many activation functions that are passed onto next layer when you input an image into a ConvNet. Main characteristics such as horizontal or diagonal edges are usually extracted by first level. This data is communicated on to the next level, which really is able to detect more complicated features like corners and combinational edges. As we go deeper into the system, it can recognise even more complex elements like objects, faces, and so on.

V. ADVANTAGES, LIMITATIONS AND APPLICATION

A. Advantages

1. High Precision
2. Low level of complexity

B. Limitations

Currently, the proposed algorithms are incapable of detecting several diseases in a single image or many instances of the same disease in a single image. It necessitates a larger data set for detection.

Conclusion

This paper presented an algorithm to hierarchically classify the tumor into three regions: whole tumor, core tumor and enhancing tumor. intensity difference, neighborhood information and wavelet features are extracted and utilized on multimodality MRI scans with various classi?ers. The use of CNN classi?er has increased the classi?cation accuracy as evident by quantitative results of our proposed method which are comparable or higher than the state of the art.

References

[1] Lina Chato, Erik Chow,” Wavelet Transform to Improve Accuracy of a Prediction Model for Overall Survival Time of Brain Tumor Patients Based On MRI Images Lina” [2018], DOI:10.1109/ICHI.2018.00091 [2] Linmin Pei, Syed M. S. Reza and Khan M. Iftekharuddin .” Improved Brain Tumor Growth Prediction and Segmentation in Longitud in al Brain MRI”,[2019] [3] G.Hemanth , M.Janardhan ,L.Sujihelen,” DESIGN AND IMPLEMENTING BRAIN TUMOR DETECTION USING MACHINE LEARNING APPROACH”[2019], DOI: 10.1109/ICOEI.2019.8862553 [4] Parveen , Amritpal singh,” Detection of Brain Tumor in MRI Images, using Combination of Fuzzy C-Means and SVM”[2015], DOI: 10.1109/SPIN.2015.7095308 [5] M. L. Goodenberger and R. B. Jenkins, “Genetics of adult glioma,” Cancer Genet., vol. 205, no. 12, pp. 613–621, Dec. 2012 [6] G. Litjens, T. Kooi, B. E. Bejnordi, A. A. A. Setio, F. Ciompi, M. Ghafoorian, J. A. W.M.vander Laak,B.van Ginneken , andC.I.S´anchez,“A survey on deep learning in medical image analysis,” Med. Image Anal., vol. 42, pp. 60–88, Dec. 2017 [7] C. Bishop, Pattern Recognition and Machine Learning. Berlin, Germany: SpringerVerlag, 2006.

Copyright

Copyright © 2022 Sumedh Sonawane, Sujit Banne, Aditya Bhor, Parag Barhate. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.