V ISVESVARAYA T ECHNOLOGICAL U NIVERSITY

Jnana Sangama Belagavi : 590018

A Phase-I Project Report on

Early Diagnosis of Diabetic Retinopathy

and Glaucoma using MATLAB

Submitted in partial fulfillment of the requirement of the award of the

Degree of

Bachelor Of Engineering
in
Electronics & Communication

Jaiganesh Rane 4NM12EC056

Keshav Padiyar M 4NM12EC065
Madhwesh M R 4NM12EC075
Ashoka Reddy V 4NM13EC404

Under the guidance of

Dr. Rekha Bhandarkar
Professor

Dept of Electronics & Communication Engineering

NMAM Institute of Technoloy
Nitte, Karkala-574110
2015-2016
 Abstract

Diabetes is a group of metabolic disease in which a person has a high blood sugar.
Diabetic Retinopathy is caused by the abnormalities in the retina due to insufficient
insulin in the body. If not diagnosed at an early stage it progresses further and the
vision of the person may deteriorate. In this work a non invasive method of approach
is implemented. So early detection of Diabetic Retinopathy is possible through image
processing using MATLAB. Hence the person can diagnose Diabetic Retinopathy at an
early stage so that it can be treated.

i
Contents

Abstract i

List of Figures vi

List of Tables vii

1 Introduction 1
1.1 Stages in Diabetic Retinopathy . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Mild Non Proliferative Retinopathy . . . . . . . . . . . . . . . 1
1.1.2 Non proliferative Retinopathy . . . . . . . . . . . . . . . . . . 1
1.1.3 Severe Non proliferative Retinopathy . . . . . . . . . . . . . . 2
1.1.4 Proliferative Retinopathy . . . . . . . . . . . . . . . . . . . . . 2
1.2 Literature Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Review of the Previous Work . . . . . . . . . . . . . . . . . . . 2
1.2.2 Present Method to Detect Diabetic Retinopathy . . . . . . . . . 4

2 Block Diagram/Flow Chart 6

2.1 Generalised Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1 Color Fundus Images . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Pre-Processing . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.4 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Data Flow Diagram of the System . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.2 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . 10

ii
3 Planning And Implementation 12
3.1 Requirement Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Time Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

References 14

iii
List of Figures

2.1 Generalised Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2 Data Flow Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

vi
List of Tables

1.1 Comparative Analysis of Performance Measures . . . . . . . . . . . . . 4

3.1 Software-Hardware Requirements . . . . . . . . . . . . . . . . . . . . 12

3.2 Project Specification details . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 Time Line Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

vii
Chapter 1

Introduction

Diabetic Retinopathy (DR) is one of the leading eye diseases that cause blindness. Au-
tomated retinal image analysis is emerging as an important screening tool for early
detection of eye diseases. Early detection of DR is necessary to reduce the risk of vi-
sion loss of diabetic patients. Retina is the light sensitive tissue (layer) lining the inner
surface of the eye. DR begins with the mild deterioration of the tiny blood vessels that
nourish the retina. Image of the retina is captured by Fundus camera, which contains
different lesions such as microaneurysms, exudates, hemorrhage and cotton wool spot.

1.1 Stages in Diabetic Retinopathy

1.1.1 Mild Non Proliferative Retinopathy

This is the earliest stage of Diabetic Retinopathy. At this stage, micro aneurysms occur
in the retina. The micro aneurysms are a tiny area of blood which is extending out from
an artery or vein in the backside of the eye.

1.1.2 Non proliferative Retinopathy

At this stage of Diabetic Retinopathy, some blood vessels that provide the nourishment
to the retina are blocked. Hence there is no proper blood circulation through the blood
vessels of the retina.

1
 CHAPTER 1

1.1.3 Severe Non proliferative Retinopathy

As the disease progresses, many blood vessels from the retina are blocked due to which
several area of retina is discontinued with blood supply. The demand for growing new
blood vessels is then send to the body through these affected areas.

1.1.4 Proliferative Retinopathy

This is the advance stage of Diabetic Retinopathy. At this stage, the growth of new
blood vessels takes place because of the signals sent by the retina for nourishment.
These newly formed blood vessels are abnormal and have thin fragile walls. If these
blood vessels leak blood, severe vision loss and even blindness can result into patient.
Color Fundus Images (CFI) taken by the medical image camera by medical personnel in
the hospital provides digitized data in the form of a fundus image that can be effectively
used for the computerized automated detection of diabetic retinopathy.

1.2 Literature Survey

This chapter intends to bring out the various motivation behind the project. Various
IEEE journals, International journals, Medical journals, text books and websites were
referred to understand the concepts involved in pre-proccesing, feature extraction from
the retinal image.

1.2.1 Review of the Previous Work

Thomas Walter et al.[1] described an algorithm for detection of exudates(a lesion in

DR eye). This algorithm helps in early diagnosis of DR. Exudates were found using
grey level variation , and their contours were determined by means of morphological
reconstruction techniques. Authors failed to distinguish between hard exudates and soft
exudates with the proposed algorithm and the evaluation of localization and distribution
of the detected exudates in order to detect macular edema.
Neera Singh and Ramesh Chandra Tripathi [2] proposed Automated Early Detection of
Diabetic Retinopathy using Image Analysis Techniques. The proposed method detects
various lesions of the retina i.e. exudates,microaneurysms and hemmorrhages and their

Dept. Of Electronics & Communication Engineering 2

 CHAPTER 1

count size and location to access the severity of the disease. Retinal Grading Algorithm
was used to check the severity of the disease.
Nikhil Amrutkar et al.[3] developed an automated image processing system which de-
tects the abnormalities in the retina using image processing techniques by applying
morphological processing techniques to Fundus image to extract the features such as
blood vessels, microaneurysms and exudates. These features were used for the detec-
tion of severity of Diabetic Retinopathy. It processed a large number of Fundus images
obtained from mass screening to help reduce the cost,increase productivity and effi-
ciency for opthalmologists.
A.J Oliveira et al.[4] proposed a method for non-proliferative DR through Fractal Di-
mension (FD). Authors used box counting FD of the macular region between normal
and mild non-proliferative DR patients and proposed this as a method of precocious di-
agnosis of Non-Proliferative Diabetic Retinopathy (NPDR). Authors failed to diagnose
NPDR using Fractal Analysis Sysem Software (TruSoft,USA).
Radim KOLAR and Jiri JAN [5] described a method for glaucomatous eye detection
based on fractal description. The fundus colour images were used, in which the areas
with retinal nerve fibres were analyzed. tthe presented method showed that FD can be
used as features or retinal nerve fibres loss detection, which is a sign of glaucomatous
eye. This is one of the risk factor for diagnosing DR. How ever accuracy of only 93.1%
was achieved by the proposed method.
In the method proposed by Meindert Niemeijer et.al [6], the green plane of the image
undergoes shade correction. The shade corrected image is top hat transformed to seg-
ment the vessels present in the image.
To completely eliminate dark lesions, a linear opening by reconstruction of size 15, a
linear closing by reconstruction of size 15, and linear opening of size 29 is done alterna-
tively. The reconstructed image is subtracted from the shade corrected image; matched
filter of size 11x11 pixels is then used. A threshold value is set which results in a binary
image. The binary image is not a replica of the original image as a result region growing
is done using the darkest pixel as the starting point.
Akara Sopharak et al. in their work to detect exudates made use of morphological fil-
tering. The pre-processing of image involves the conversion in RGB space to Hyper
spectral imaging (HSI) colour space, as the intensity components are removed from

Dept. Of Electronics & Communication Engineering 3

 CHAPTER 1

Table 1.1: Comparative Analysis of Performance Measures

Techniques used for Techniques used for
Feature Extracted Performance Measured
pre-processing feature extraction
Sensitivity-100%
Lesions Median filtering Morphological Filtering
Accuracy-87%
Sensitivity-85.4%
Microaneurysms Median filtering Double ring filter
Accuracy-83.1%
Sensitivity-80%
Exudates Histogram Equalization Fuzzy C-means Clustering
Specificity-99.5%
Morphological Sensitivity-94.78%
Exudates Neural Network System
Transformation Specificity-94.29%
Microaneurysms Thresholding Double ring filter Sensitivity-68%
Sensitivity-87.28%,
Accuracy-99.11
Morphological
Exudates Fuzzy C-means Clustering Specificity-99.24%
Reconstruction
PPV-42.77%
PLR- 224.26

the remaining colour components.Contrast adaptive histogram equalization is applied

to small regions of the image. The small regions are combined by bilinear interpolation
after equalization. Exudates are regions of high intensity values[7].
Table 1.1 shows the performance measures of various algorithms for diagnosis of
DR[8].

1.2.2 Present Method to Detect Diabetic Retinopathy

Angiography

Angiography is the X-ray(radiographic) study of blood vessels. Angiography uses ra-

diopaque substance or contrast medium to make the blood vessels visible under X-ray.
The key ingredient of radiographic contrast medium is iodine. Angiography technique
is used to find various blood vessels related disorders.

Fluorecein Angiography

It is the angiographic technique for monitoring the eye related disorders. It is an eye test
that uses special dye and camera to look into the blood flow in the retina and choird re-
gion of eye. It involves injection of sodium fluorescein(contrast medium into the body)
and then an angiogram is obtained by photographing the fluorescence emitted after illu-
mination of retina with blue light. This is the modern day technique used for diagnosis

Dept. Of Electronics & Communication Engineering 4

 CHAPTER 1

of DR. But this technique has got drawbacks. Some of the drawbacks as follows:
1. Nausea, vomiting, gastrointestinal distress, headache can occur due to angiography.
2. Patients who have blood-clotting problems have allergy to the contrast medium can-
not undergo this technique.
3. X-rays have risk of ionising radiation exposure.
4. People with kidney disease or injury may suffer further damage from the contrast
medium.

By using the above mentioned techniques of preprocessing and feature extraction,

the best technique which gives the desired result with better performance measures is
approached.

Dept. Of Electronics & Communication Engineering 5

Chapter 2

Block Diagram/Flow Chart

2.1 Generalised Block Diagram

Fig. 2.1 shows the generalised block diagram for processing the fundus retinal im-
age.Description of the blocks are given below:

2.1.1 Color Fundus Images

Colored images of eye are captured through Fundus Camera. A fundus camera is a
specialized low power microscope with an attached camera. Though fundus camera
images retinal,microaneurysms and blood vessels of eye are easier to visualize. Sharp
focused fundus camera image is used as the input image and is processed futher.

2.1.2 Pre-Processing

These images are preprocessed to improve the quality of image and this preprocessed
image is used for the further stages. A suitable preprocessing technique is used so that
a particular feature can extracted from the image. various preprocessing techniques are
Median filtering, Morphological Transformation, Thresholding, Histogram Equalisa-

Figure 2.1: Generalised Block Diagram

6
 CHAPTER 2

tion etc.

2.1.3 Feature Extraction

The technique of extracting features from an image which contains relevant features,
so that the desired task can be performed. The technique used for feature extraction
depends on type of the feature extracted. The various feature extracting techniques are
Morphological Filtering, Double ring filter, Neural Network System etc.

2.1.4 Classification

Based on the results obtained from the feature extraction, image is classified into a
particular category.

2.2 Data Flow Diagram of the System

Figure 2.2 shows the data flow diagram of the system. The retinal image captured
through Fundus camera is read as input. The input image is preprocessed to obtain 2D
image (Red, Green, Blue channel images). Normalization or De-noising of the image
is performed to remove the unwanted noise and enhance the image quality. Feature
Extraction includes the methods like Median filtering, Morphological Transformation,
Thresholding, Histogram Equalisation etc. By using the best estimated method, features
like microaneurysms, lesions are extracted. Based on the results of feature extraction, a
decision is made whether the patient is prone to DR or not.

2.2.1 Preprocessing

Various techniques used for preprocessing are discussed below:

Median Filtering

The green plane image is shade corrected image is applied using a 25x25 median filter.
The size of the median filter was chosen such that it is wider than the widest blood
vessel in used dataset of images. The shade corrected image has negative values for all
pixels which have intensity lower than the background. By removing all pixels with a

Dept. Of Electronics & Communication Engineering 7

 CHAPTER 2

Figure 2.2: Data Flow Diagram

Dept. Of Electronics & Communication Engineering 8

 CHAPTER 2

positive value from the image, bright pathologies no longer influence the later analysis.
All pixels with a positive value were set to zero resulting in the preprocessed image. The
green plane of the image (I) is filtered using a 3x 3 median filter to remove occasional
salt-and-pepper noise[9].
The green plane image is shade corrected image is applied using a 56x56 median filter
[10]. A top hat transform is applied using a 25 pixel long linear structuring element
which is spaced 180 degrees apart. Two matched filters are used to approximate the
characteristics of the objects of different sizes and each is threshold and combined. The
result is a micro aneurysm seed image. The region growing process is carried out using
the seed pixel from which eighty features are extracted. A series of image processing
techniques were used in the method proposed by Atushi Mizutani et al [18] which
includes brightness correction, gamma correction and contrast enhancement. Median
filtering is employed to reduce the effect of noise.

Morphological Transformation

Pre-processing is carried out for detection of Dark lesions which have very low content
in the green colour plane[11]. Moreover the background intensity variation of the image
in the green plane is less. Hence the green channel of an image is obtained initially and
for every pixel p in the image, a neighborhood is considered centered on the selected
pixel. If the gray level of the pixel is lesser than the fraction of the mean of then the
pixel is considered as darker than the surrounding pixels. The darker areas than the
normal retina are detected. Finally a binary image is obtained with non zero values for
selected dark pixels. In the method proposed by Meindert Niemeijer et.al[6], the green
plane of the image undergoes shade correction. The shade corrected image is top hat
transformed to segment the vessels present in the image.
Morphological opening with a linear structuring element of size 15 pixels is used to
remove rounded bright zones of size less than 15 pixels. The contrast is improved
by obtaining the sum of the top hat transforms. The resulting images are noisy and
Gaussian smoothing is used to eliminate the noise present in the image. To completely
eliminate dark lesions, a linear opening by reconstruction of size 15, a linear closing
by reconstruction of size 15, and linear opening of size 29 is done alternatively. The
reconstructed image is subtracted from the shade corrected image; matched filter of size

Dept. Of Electronics & Communication Engineering 9

 CHAPTER 2

11x11 pixels is then used. A threshold value is set which results in a binary image.

Histogram Equalization

A neural network based preprocessing stage for histogram equalization and specializa-
tion is utilized [7]. A sliding window technique uses the regions covered by the window
serve as the input features. For preprocessing, the green plane shaded colour images are
considered. If the exudates are the center pixel then it is considered as positive sample
else it is considered as negative sample. To locate the exudates, a 9x9 sized window
is used. A target vector is created based on the positive and negative samples and the
vector is set as 0 or 1.

Thresholding

S.Kavitha et al.[14] proposed a method for detection of hard and soft exudates wherein
the preprocessing involves colour space conversion into lab colour space and detection
of fundus region. Further the Lab colour space image is converted into binary image by
thresholding. The binary image is morphologically closed and a fundus mask is created
with pixels at the fundus marked as 1 and pixels in the background marked as 0.

2.2.2 Feature Extraction

Various techniques used for feature extraction are discussed below:

Double Ring Filter

After image pre-processing, candidate regions for microaneurysms were detected using
a double-ring filter [15]. The diameters of 5 and 13 pixels for inner and outer ring re-
spectively is used for detecting micro aneurysms in the surrounding retinal regions of
the image. Any potential false positives located in the regions corresponding to blood
vessels were removed by automatic extraction of blood vessels from the images. One
hundred twenty six image features were determined and 28 components were selected
by using principal component analysis. The candidate lesions were classified into mi-
croaneurysms or false positives using the rule-based method and an artificial neural
network.

Dept. Of Electronics & Communication Engineering 10

 CHAPTER 2

Morphological Filtering

Morphological closing leads to twisted vessel like patterns being treated as DR, an ideal
operator with all structuring elements having large diameter is used [16]. The binary
features, gray level and colour features are extracted for the detection of DR.

Density Function

The spatial relationship of the pixels has considered to differentiate between the can-
didate lesions, noise and pigment variability. A density function is defined to find the
number of similar pixels in the neighborhood of the selected pixel. A threshold value is
set to eliminate pixels belonging to the optic disk and the blood vessels as a result the
dark lesions are identified [17].

Region Growing

Region growing algorithm is used to find the morphology of the candidate. The fea-
tures are defined to further improve the accuracy of detection in the method [?]. A
modification is made so that the measures were multiplied with a scale factor for better
discrimination between micro aneurysms and other parts of the vasculature.

Dept. Of Electronics & Communication Engineering 11

Chapter 3

Planning And Implementation

3.1 Requirement Analysis

Table 3.1 shows the hardware and software requirements for the project. These are the
minimum requirements and the higher versions are entertained.

Project Specifications

According to the previously obtained results using the techniques, which will be used
in this work, the estimated output is shown in table 3.2

Project Deliverables

This Section should contain the outputs that will be delivered at the end of your project

Major Component List

This Section should contain the Major Component List and other requirements of your
project

Table 3.1: Software-Hardware Requirements

Software Requirements Hardware Requirements
MATLAB (version 7.10.0.499) 500GB HDD with i5 Processor
OS (Windows 7) 4GB RAM

12
 CHAPTER 5

Table 3.2: Project Specification details

Sensitivity Specificity Accuracy
100% 90% 98%

Table 3.3: Time Line Details

Sl no. Name USN Description Duration
1. Jaiganesh Rane 4NM12EC056 Feature Extraction 00/00/2016
2. Keshav Padiyar M 4NM12EC065 Pre-processing 00/00/2016
3. Madhwesh M R 4NM12EC075 Pre-processing 00/00/2016
4. Ashoka Reddy V 4NM13EC404 Feature Extraction 00/00/2016
Interfacing both
5. Group Work - 00/00/2016
modules and testing

3.2 Time Lines

Table 3.3 shows the modules assigned to each member with time line details.

Dept. Of Electronics & Communication Engineering 13

References

[1] Thomas Walter, Jean Claude Klein, Pascale Massin and Ali Erginay, A Contri-
bution of Image Processing to the Diagnosis Diabetic Retinopathy Detection of
Exudates in Color Fundus Images of the Human Retina, IEEE Transactions on
Medical Imaging, Vol. 21, No.10, pp. 0278-0062,2002.

[2] Neera Singh, Ramesh Chandra Tripathi Automated Early Detection of Diabetic
Retinopathy using Image Analysis Techniques, International Journal of Com-
puter Applications, vol. 8.,no. 2 ,pp. 0975-8887,October 2010.

[3] Nikhil Amrutkar, Yogesh Bandgar, Sharad Chitalkar and S.L. Tade.Retinal Blood
Vessel Segmentation Algorithm for Diabetic Retinopathy and Abnormality De-
tection using Image Substraction, International Journal of Advanced Research in
Electrical, Electronics and Instrumentation Engineering, vol. 2,Issue 4,pp. 1547-
1553, April 2013.

[4] A.J. Oliveira, A.C.B. Kunicki, M.B.M. Mendonca,C.T.F. Barbosa,Can the Frac-
tal Dimension be Applied for the Early Diagnosis of Non-Proliferative Dia-
betic Retinopathy?,Brazilian Journal of Medical and Biological Research, vol.42,
no.10, pp. 930-934,2009.

[5] Radim KOLAR, Jiri JAN,Detection of Glaucomatous Eye via Color Fundus Im-
ages using Fractal Dimensions,Radio Engineering,vol.17,no.3,2008.

[6] Meindert Niemeijer, Bram van Ginneken, Joes Staal, Maria S. A.Suttorp-Schulten,
and Michael D. Abrmoff, ?Automatic Detection of Red Lesions in Digital Color
Fundus Photographs?, IEEE transactions on medical Engineering, vol. 24, no. 5,
May 2005.

14
 CHAPTER 5

[7] Akara Sopharak, Bunyarit Uyyanonvara , Sarah Barmanb, Thomas H. Williamson,

Automatic detection of diabetic retinopathy exudates from non dilated retinal im-
ages using mathematical morphology methods,Computerised Medical Imaging
and Graphics,Vol.32,pp 720-727,2008.

[8] A.P.Shingade, A.R.Kasetwar,A Review on Implementation of Algorithms for De-

tection of Diabetic Retinopathy,International Journal of Research in Engineering
and Technology,vol. 03,pp. 87-93,March 2014.

[9] Alan D. Fleming, Sam Philip, Keith A. Goatman, John A. Olson, and Peter
F,Automated Microaneurysm Detection Using Local Contrast Normalization and
Local Vessel Detection?, Sharp IEEE Transactions on medical imaging, vol. 25,
no. 9, September 2006.

[10] Harihar Narasimha-Iyer, Ali Can, Badrinath Roysam, Howard L. Tanenbaum,

and Anna Majerovics, Integrated Analysis of Vascular and Nonvascular Changes
from Color Retinal Fundus Image Sequences, IEEE Transactions on biomedical
engineering, vol. 54, no. 8, August 2007.

[11] Kenneth W. Tobin, Edward Chaum, V. Priya Govindasamy, and Thomas P.

Karnowski,Detection of Anatomic Structures in Human Retinal Imagery?, IEEE
Transactions on medical imaging, vol. 26, no. 12, December 2007.

[12] Gerald Schaefer and Edmond Leung, An investigation into neural network for
detection of exudates in retinal images?, Applications of Soft Computing, ASC
52, pp. 169 177.2009.

[13] IRF540 Data Sheet, International Rectifiers Inc, 2003.

[14] S.Kavitha and K.Duraiswamy,?Automatic detection of hard and soft exudates in

fundus images using colour histogram thresholding?,European Journal of Scien-
tific Reasearch,Vol.48,No.3,pp.493-504,2011.

[15] Atsushi Mizutani,Chisako Muramatsu,Yuji Hatanaka,? Automated Microa-

neurysm detection method based on double-ring filter in retinal fundus images?
IEEE Transactions on Medical imaging , Proc. of SPIE Vol. 7260, 72601N
15,2009.

Dept. Of Electronics & Communication Engineering 15

 REFERENCES

[16] Diego Marn, Arturo Aquino, Manuel Emilio GegndezArias, and Jos Manuel
Bravo ?A New Supervised Method for Blood Vessel Segmentation in Retinal Im-
ages by Using GrayLevel and Moment Invariants-Based Features?, IEEE Trans-
actions on medical imaging, vol. 30,no. 1, January 2011.

[17] Xiayu Xu, Meindert Niemeijer, Qi Song, Milan Sonka et.al., ?Vessel Boundary
Delineation on Fundus Images Using Graph-Based Approach?, IEEE Transac-
tions on medical imaging, vol. 30, no. 6, June,2011. bibitemAAH Aliaa Abdel-
Haleim Abdel-Razik Youssif, Atef Zaki Ghalwash, and Amr Ahmed Sabry Abdel-
Rahman Ghoneim, Optic Disc Detection From Normalized Digital Fundus Im-
ages by Means of a Vessels Direction Matched Filter?, , IEEE Transactions on
medical imaging, vol. 27, no.1, January 2008.

Dept. Of Electronics & Communication Engineering, NMAMIT 16