Brain tumor detection by scanning MRI images (using filtering techniques)

1
of
36
REVIEW – 01 PRESENTATION FOR DIGITAL IMAGE PROCESSING (SWE1010)
Brain Tumor Detection by Scanning
MRI Images
(Using Filtering Techniques)
-by
15MIS0144 R Vivek

2
of
36
Problem Statement
In today’s world,
one of the reason in rise of mortality
among the people is BRAIN CANCER.
For detection of brain tumor, first we have to read the MRI
image of brain and then we can apply segmentation of the image.
Here we present an Efficient method for removing noise from
the MRI image as well as for brain tumor detection.

3
of
36
What is Brain Tumour?
 A tumor can be defined as any mass caused by abnormal or
uncontrolled growth of cells. This mass of tumor grows within
the skull, due to which normal brain activity is hampered.
 Which if not detected in earlier
stage, can take away the
person’s life. Hence, it is very
important to detect the brain
tumor as early as possible.

4
of
36
Motivation
The Motivation for doing this project is primarily an interest of
undertaking one of the best challenging j-component project in
an interesting area of research i.e., Brain tumor detection.
We hope that this is the opportunity to learn about a new area of
computing not covered in lectures was appealing. This brain
tumor detection is possibly an area that we might’ve worked
upon, if we were in that stream. But now we’ve luckily got a
chance to show-up our self by this j-component.

5
of
36
Introduction
Image Segmentation is the process of partitioning a digital
image into multiple regions or sets of pixels which are similar
with respect to some characteristic such as colour, texture or
intensity.
Adjacent regions are significantly different with respect to the
same characteristics.
Segmentation produces a set of non-overlapping regions whose
union is the entire image.
Segmentation algorithms for images generally based on the
discontinuity and similarity of image intensity values.

6
of
36
Introduction (cont.,)
 Therefore the choice of image segmentation technique is
problem dependent that has been considered. So in our project,
we made an attempt to pick up already segmented images
(which are more over much noise free) and further
smoothening the image using various types of filters and
analyse their effectiveness.
 Based upon, the result obtained by processing the image
through different filters, it is clear visible for a normal naked
human eye to detect the tumour in the image.

7
of
36
Objective
Our Main Objective of medical imaging of BRAIN TUMOR
- is to extract meaningful and accurate information from
these images with least error possible and
 Finally conclude whether it’s a tumor image or not.

8
of
36
Our Proposed Method
-Framework
Load MRI Brain image sequence
Pre-processing
Tumor Segmentation using
different filters
Tumor identification based
on grade
Water
shed
Filter
Dilation
Filter
Erosion
Filter
Thresh
old
Filter
Median
Filter
Grey
Filter

9
of
36
Literature Survey
Reference Paper - 01
A Noble Approach for Noise removal from Brain
Image using region filling
 Firstly, they started with image acquisition, then noise is
removed from the noisy image.
 According to them, noise means in MRI images there will be
some information regarding the institute etc.,
 So all that unnecessary information which is treated as noise will
be removed.
 Then they converted RGB to Grey scale image.
 After this, they started to apply region filling, means they
selected some particular region of interest, then they filled that
area.
 Finally applied some low range filters.

10
of
36
Literature Survey

11
of
36
Literature Survey
Automatic Segmentation framework for primary tumors
from Brain MRI’s using Morphological Filtering techniques.
 Used a weighted algorithm.
 Firstly to the image identified, the intensity is adjusted
 Then they performed Morphological Erosion & Morphological
dilation.
 This process is further followed by Image subtraction, then
they found the threshold value by making use of some
histogram techniques.
 thus they performed Binary thresholding continued by
morphological labelling.
 Finally the segmented tumor is processed or know clearly
by image masking.

12
of
36
Literature Survey
Segmentation
process
Segmentation
results

13
of
36
Literature Survey
Morphological based segmentation of
Brain Image for tumor detection
Three major phases, namely,
1. pre-processing, (includes steps like
converting it to gray from a colour image
followed by type-casting the image)
2. Image segmentation, (followed the concept
of Thresholding technique and edge based
detection, and some operators related to
that are applied)
3. Image post processing. (Using filters)

14
of
36
MRI Brain Tumour segmentation with region
growing method based on the gradients
and variances along and inside of the
boundary curve.
Based on the gradients and variances along and inside
of the boundary curve, which focuses on the
consistency of the region and the smoothness of
the boundary.
Finally, what I observed is that, they made it easy to
solve the problem of threshold selection, i.e., the
minimum of the function value is the optimum
result which corresponding to the desirable
threshold.
Literature Survey

15
of
36
A Survey on brain tumor detection using Image Processing Techniques
Followed a four step categorical processes in order to detect the brain tumour from
MRI images.
Those four different categories are
Pre-processing,
Segmentation,
Optimization,
and feature extraction.
Literature Survey
Also they made some research
by reviewing other papers done
by professionals, and explained
by listing out various techniques
in use and also a brief description
is explained by them.

16
of
36
A survey on detection of brain tumor from MRI Brain images
Done Some research on Brain tumor related IEEE papers, and they explained
the flaws limitations along with the accuracy.
K nearest neighbours (KNN) and conventional fuzzy connected c-mean
(FCM), [This is the best method with 100% accuracy].
Some other methods with 100% accuracy are wavelet entropy approach,
neural network based method, feature extraction using PCA (principal
component analysis) and LDA (Linear Discriminant analysis) and by using
SVM (Support vector machines) classifiers.
Literature Survey

17
of
36
An Adaptive Filtering technique for brain tumor analysis and
detection
 They put forth a method for detection and segmentation of the tumor.
 The method is a segmentation process of 2D MRI image using various
filtering techniques. MATLAB has been used for the implementation.
 The idea which they’ve used for detection of brain tumour actually
grabbed my attention and this is the method that I’m going to work on for
this DIP project.
 They’ve used certain filters, for edge detection, image
sharpening and image enhancement.
Literature Survey

18
of
36
Literature Survey

19
of
36
Brain tumor detection and segmentation using conditional
random field
Followed mainly three steps for Tumor detection:
Initial segmentation,
modelling of energy function
& optimize the energy.
In their framework they incorporated additional information present in
weighted MRI images, which made the performance better in presence of
artifacts and helps to improve boundaries.
Literature Survey

20
of
36
Literature Survey

21
of
36
Efficient detection of brain tumor from MRIs using K-means
segmentation and normalized histogram
They used many image de-noising filters such as Median filter, Adaptive
filter, Averaging filter, Un-sharp masking filter and Gaussian filter are
used to remove the additive noises present in the MRI images i.e.,
Gaussian, salt & pepper noise and speckle noise.
Literature Survey

22
of
36
Literature Survey

23
of
36
Brain tumor detection in conventional MRI images based
on Statistical texture and morphological features
Followed four steps for brain tumor diagnosis:
Image preprocessing,
Tumor segmentation,
Selected features extraction,
Automatic tumor grade
identification using classifiers.
Literature Survey

24
of
36
Comparative study on brain tumor detection techniques
Two parts:
1. Pre-processing (used local binary pattern)
2. Segmentation (used different techniques like edge detection and
morphological operation like erosion and dilation).
Literature Survey

25
of
36
Literature Survey

26
of
36
An efficient Brain tumor dection from MRi images using entropy
measures
 They are comparing & analysing various threshold entropy based
segmentation methods on the basis of simulation results.
 Entropy methods are applied to the MRI images of brain tumor or any
internal structure of our body, are compared & analysed.
 Based upon their comparing and analysing through simulation results, they
observed that Havrda-charvat entropy performs better than any other
entropy algorithms.
Literature Survey

27
of
36
Literature Survey

28
of
36
A survey on brain tumour detection using image processing
techniques:
 Done some survey about the brain tumour detection by referring to
different IEEE papers.
 Mainly divided into three steps:
pre-processing,
segmentation,
post-processing.
 Different segmentation techniques mainly used
threshold based segmentation
Region based
fuzzy c-means ,k-means etc.,
Literature Survey

29
of
36
Brain tumour pixels detection using adaptive wavelet based
histogram thresholding and fine windowing
METHODOLOGY:
 pre-processing
 Contrast stretching applied for enhancement of image after converted to
grey level image.
 Two-level wavelet applied horizontal, vertically and stored
for Future processing.
 Thresholding
-global thresholding
-local thresholding
-window thresholding
Literature Survey

30
of
36
NORMAL MRI IMAGE:
TUMOUR contains MRI IMAGE:

31
of
36
Brain tumour detection based on watershed transformation
This paper identifies the tumour accurately by following this steps:
Pre-processing ,
watershed,
Transformation,
Threshold,
Morphological,
Background-marker,
segmented-output
Literature Survey

32
of
36
Literature Survey

33
of
36
Detection of brain tumour using NNE Approach:
This paper proposes a methodology with 4 steps to identify the brain tumor:
pre-processing ,
segmentation ,
feature extraction,
classification.
classification done by using
Neural network.
Literature Survey

34
of
36
Brain image(without tumor):
Normal image pre-processed image canny-edge segment resultant(NNI)(no
tumor)
Brain image(with tumor):
MRI image pre-processed segmentation tumor detection

35
of
36
Brain tumour detection in MRI images using PNN and GRNN:
Tumour segmentation for MRI brain images:
-- K-means clustering
--Fuzzy c-mean clustering.
Classification is done by using
--Probability neural network
--Generalized regression neural network.
Original image pre-processed apply k-means clustering.
Literature Survey

36
of
36
Detection of a brain tumour using segmentation and
morphological operators from MRI scan with FPGA:
This paper proposes a method to detect brain
Tumour by:
---segmentation
---Morphological operators.
Literature Survey

37
of
36
Brain tumour diagnosis from MRI feature analysis – A
comparative case study:
 This is an comparative study of transform techniques namely
--- Discrete cosine Transform.
---Discrete Wavelet transform.
 Each transform technique is applied separately.
 Later feature Extraction and classification of tumor in MRI image is done.
Literature Survey

38
of
36
Automatic Brain tumour tissue detection based on hierarchical
centroid shape descriptor in T1-weighted MR images
Proposed methodology is used to detect the tumour in
poor-contrast images
irregular shape-tumour.
Methodology :
1.Brain Extraction (original image)
2.k-means method
3.Thresolding
4.HCSD(hierarchical centroid shape descriptor)
5.co-ordinating for boundary box
6.Superimposing original image with 5step.
Literature Survey

39
of
36
Hybrid Approach for brain tumour detection and classification in
Magnetic resonance images:
This research paper is about detecting brain tumor by using Hybrid approach.
 HYBRID APPROACH = region based + texture based methods.
 Methodology includes five steps:
 Segmentation is done by using FBB algorithm.
Literature Survey

40
of
36
A novel methodology for brain tumor detection based on two
stage segmentation of MRI images
 Two-stage segmentation
 Apply Gabor filter
 Contour level segmentation method.
Input images:
Literature Survey

42
of
36
Automatic detection, extraction and mapping of brain tumor from
MRI images using frequency emphasis homomorphic and
cascaded hybrid filtering techniques:
Using homomorphic filtering
Noise removed by Gaussian method algorithms
Hybrid filters used to remove domain noises.
This not only detect tumour region but also
point exact position in brain image.
LIMITATION:
•Using Butterworth high pass filter instead of
Gaussian high pass filter with homomorphic filter
Works better.
Literature Survey

43
of
36
Watershed segmentation brain tumour detection:
 Uses watershed algorithm for segmentation.
 Uses median filter and bilateral filters to remove
noise in MRI images.
Literature Survey

44
of
36
A hybrid approach for detection of brain tumour in MRI images:
 Detect brain tumor in MRI images by combining Classification and
clustering algorithms.
 Which decreases complexity of time and memory.
 Phase-1
----- Non-matrix factorization with sparseness constraint method used to
separate ROI from image.
 Phase-2
------classification of ROI is performed (using top-LBP )
Literature Survey

45
of
36
Literature Survey - Inference

46
of
36
Proposed Architecture
(Future work)
Load MRI Brain image sequence
Pre-processing
Tumor Segmentation using
different filters
Tumor identification based
on grade
Water
shed
Filter
Dilation
Filter
Erosion
Filter
Thresh
old
Filter
Median
Filter
Grey
Filter

47
of
36
Simulation Tool
Matlab simulation tool is used for coding the algorithm & for designing my
application’s GUI.
Our project
progress

48
of
36
Dataset Description
We’ve created our own database by downloading some brain tumour images from
online datasets.

49
of
36
References
[1.] Daizy Deb, Bahnishikha Dutta and Sudipra Roy “A Noble Approach for Removal from Brain Image using Region Filling”, IEEE International
Conference on Advanced Communication Control and Computing Technologies, 2014.
[2.] Resmi S. Ananda and Tessamma Thomas “Automatic segmentation framework for primary tumors from brain MRIs using morphological filtering
techniques”, 5th
International Conference on BioMedical Engineering and Informatics, 2012.
[3.] Amlan Jyoti, Mihir Narayan Mohanty and Mallick Pradeep Kumaar “Morphological Based Segmentation of brain Image for tumor detection”,
International Conference on Electronics and Communication Systems, 2014.
[4.] Wankai Deng, Wei Xiao and Jianguo Liu “MRI brain tumor segmentation with region growing method based on the gradients and variances along
and inside of the boundary curve”, 3rd
International Conference on Biomedical Engineering and Informatics, volume 1, 2010.
[5.] Luxit Kapoor and Sanjeev Thakur “A survey on brain tumor detection using Image processing techniques”, 7th
International Conference on Cloud
Computing, Data Science & Engineering – Confluence, 2017.
[6.] S. U. Aswathy, G. Glan Deva Dhas and S. S. Kumar “A survey on detection of brain tumour from MRI Brain images”, 7th
International Conference on
Cloud Computing, Data Science & Engineering –Confluence, 2017.
[7.] Minu Samantaray, Millee Panigrahi, K. C. Patra, Avipsa S. Panda and Rina Mahakud “An adaptive filtering technique for brain tumor analysis and
detection”, 10th
International Conference on Intelligent and Control (ISCO), 2016.
[8.] C. Hemasundara Rao, P. V. Naganjaneyulu and K. Satya Prasad “Brain tumour detection and segmentation using conditional random field”, IEEE 7th
International Advance Computing Conference (IACC), 2017.
[9.] Garima Singh and M. A. Ansari “Efficient detection of brain tumor from MRIs using K-means segmentation and normalized histogram”, 1st
India
International Conference on Information Processing (IICIP), 2016.

50
of
36
References (cont.,)
[10.] Manu Gupta, B. V. V. S. N. Prabhakar Rao and Venkateswaran Rajagopalan “Brain tumour detection in conventional MR images based on
statistical texture and morphological features”, International Conference on Information Technology (ICIT), 2016.
[11.] D. Haritha “Comparative study on Brain tumor detection techniques”, International Conference on Signal Processing, Communication, Power
and Embedded System (SCOPES), 2016.
[12.] Devendra Somwanshi, Ashutosh Kumar, Pratima Sharma and Deepika Joshi “An efficient Brain Tumor Detection from MRI Images Using
Entropy Measures”, International Conference on Recent Advances and Innovations in Engineering (ICRAIE), 2016.
[13.] Luxit Kapoor and Sanjeev Thakur “A survey on brain tumor detection using image processing techniques”, 7th
Internationa; Conference on
Cloud Computing, Data Science & Engineering – Confluence, 2017.
[14.] Sanjivani Salwe, Ranjana Raul and Pratik Hajare “Brain Tumor Pixels detection using adaptive wavelet based histogram thresholding and fine
windowing”, International Conference on Information Technology (InCITe) – The Next Generation IT Summit on the Theme – Internet of
Things: Connect your Worlds, 2016.
[15.] K. Ramya and L. K. Joshila Grace “Brain tumour detection based on watershed transformation”, Interational Conference on Communication
and Signal Processing (ICCSP), 2016.
[16.] Kanwarpreet Kaur, Gurjot Kaur and Jaspreet Kaur “Detection of brain tumour using NNE Approach”, IEEE International Conference on
Recent Trends in Electronics, Information & communication Technology (RTEICT), 2016.
[17.] K. S. Thara and K. Jasmine “Brain tumour detection in MRI images using PNN and GRNN”, International Conference on Wireless
Communications, Signal Processing and Networking (WiSPNET), 2016.

51
of
36
References (cont.,)
[18.] H M Willian Thomas and S C Prasanna Kumar “Detection of a brain tumor using segmentation and morphological operators from MRI scan
with FPGA”, International Conference on Applied and Theoretical Computing and Communication Technology (iCATccT), 2015.
[19.] Shobana G and Ranjith Balakrishnan “Brain tumor diagnosis from MRI feature analysis”, 2015 International Conference on Innovations in
Information, Embedded and Communication systems (ICIIECS), 2015
[20.] Elisee LLunga-Mbuyamba, Juan Gabriel Avina-Cenvantes, Dirk Lindner, Jesus Guerrero-Turrubiates and Claire Chalopin “Automatic Brain
tumour tissue detection based on hierarchical centroid shape descriptor in T1-weighted MR images”, International Conference on Electronics,
Communications and Computers (CONIELECOMP), 2016.
[21.] Praveen G. B., Anita Agrawal “Hybrid Approach for brain tumour detection and classification in Magnetic resonance images”, 2015
Communication, Control and Intelliegent Systems (CCIS), 2015.
[22.] Anjali Joshi, V. Charan and Shanthi Prince “A novel methodoloy for brain tumor detection based on two stage segmentation of MRI images”,
2015 International Conference on Advanced Computing and Communicatiob Systems, 2015.
[23.] Rana Banik, Md. Rabiul Hasan and Md. Saif Iftekhar “Automatic detection, extraction and mapping of brain tumor from mri images using
frequency emphsis homomorphic and cascaded hybrid filtering techniques”, 2015 International Conference on Electrical Engineering and
Information Communication Technology (ICEEICT), 2015.
[24.] Padmakant Dhange, M. R. Phegade and S. K. Shah “Watershed segmentation brain tumor detection”, 2015 International Conference on
Pervasive Computing (ICPC), 2015.
[25.] Solmaz Abbasi and Farshad Tajeri Pour “A hybrid Approach for detection of brain tumor in MRI images”, 2014 21th
Iranian Conference on
Biomedical Engineering (ICBME), 2014.

Brain tumor detection by scanning MRI images (using filtering techniques)

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (20)

Similaire à Brain tumor detection by scanning MRI images (using filtering techniques)

Similaire à Brain tumor detection by scanning MRI images (using filtering techniques) (20)

Dernier

Dernier (20)

Brain tumor detection by scanning MRI images (using filtering techniques)