Diabetic retinopathy screening. Stages of diabetic retinopathy.
Detection of Diabetic Retinopathy Diseases using …...2019/09/07 · is called its 4-neighbors....
Transcript of Detection of Diabetic Retinopathy Diseases using …...2019/09/07 · is called its 4-neighbors....
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
123 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
Detection of Diabetic Retinopathy Diseases using Neural Network
Manmohan Sharma1, Er. Sandeep Singh Dhaliwal
2
1P.G. Student, Department of Computer Engineering, BMS Engineering College, Sri Muktsar Sahib, India 2Head of Deptt. Assit. Prof. CSE in Bhai Maha Singh College of Engineering, Sri Muktsar Sahib, Punjab
Abstract:- Here we address the detection of Hemorrhages and micro aneurysms in color fundus images. In pre-Processing we separate red,
green, blue color channel from the retinal images. The green channel will pass to the further process. The green color plane was used in the
analysis since it shows the best contrast between the vessels and the background retina. Then we extract the GLCM (Gray Level Co-Occurrence
Matrix) feature. In the GLCMs, several statistics information is derived using the different formulas. These statistics provide information about
the texture of an image. Such as Energy, Entropy, Dissimilarity, Contrast, Inverse difference, correlation Homogeneity, Auto correlation,
Cluster Shade Cluster Prominence, Maximum probability, Sum of Squares will be calculated for texture image. After feature Extraction, we
provide this feature to classifier. Finally it will predict about the retinal whether it is hemorrhages or micro-aneurysms . After predicting the
about the retinal image we will localize the affected place. For segmenting the localized place we will use adaptive thresholding segmentation.
__________________________________________________*****_________________________________________________
I. INTRODUCTION
Diabetic retinopathy (DR) is a vascular disease of the retina
which affects patients with diabetes mellitus. It is the
number one cause of blindness in people between the ages
of 20-64 in the United States. It is, therefore, a worthwhile
topic for all medical students to review. Diabetes mellitus is
extremely common, so it is not surprising that DR affects
3.4 percent of the population (4.1 million individuals). Of
the millions of people with DR, nearly one-fourth have
vision-threatening disease (AAO 2008). The likelihood of
developing diabetic retinopathy is related to the duration of
the disease. Type 2 diabetes has an insidious onset and can
go unnoticed for years. As a result, patients may already
have DR at the time of diagnosis. Type 1 diabetics, on the
other hand, are diagnosed early in the course of their
disease, and they typically do not develop retinopathy until
years after the diagnosis is made. The risk of developing
retinopathy increases after puberty. Twenty years after the
diagnosis of diabetes, 80% of type 2 diabetics and nearly all
type 1 diabetics show some signs of retinopathy (Klein
1984a, Klein 1984b). While these numbers are eye-opening,
diabetics can decrease their risk of retinopathy and slow the
progression of the disease after it has begun with tight
glucose control (DCCTRG 1993).
Figure 1: Normal Fundus
Figure 2: Diagram of Normal Eye
Anatomy
The retina is a multi-layered sheet composed of neurons,
photoreceptors, and support cells. It is one of the most
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
124 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
metabolically active organs in the body, and as a result, it is
extremely sensitive to ischemia and nutrient imbalances
(Frank 2004). A perfused retina is a happy retina.
Classification
NPDR
Symptoms
Screening
Examination
Treatment Options
Images and pictures
To formulate the adjacency criterion for connectivity, we
first introduce the notation of neighborhood. For a pixel p
with the coordinates (x,y) the set of pixels given by:
is called its 4-neighbors. Its 8-neighbors are defined as
From this we can infer the definition for 4- and 8-
connectivity:
Two pixels p and q, both having values from a set V are 4-
connected if q is from the set and 8-connected if q is
from .
A set of pixels in an image which are all connected to each
other is called a connected component. Finding all
connected components in an image and marking each of
them with a distinctive label is called connected component
labeling. An example of a binary image with two connected
components which are based on 4-connectivity can be seen
in Figure 1. If the connectivity were based on 8-neighbors,
the two connected components would merge into one.
Figure 3 :Two connected components based on 4-
connectivity.
Some Applications:
Image processing has an enormous range of applications;
almost every area of science and technology can make use
of image processing methods. Here is a short list just to give
some indication of the range of image processing
applications.
1. Medicine
• Inspection and interpretation of images obtained
from X-rays, MRI or CAT scans,
analysis of cell images, of chromosome
karyotypes.
2. Agriculture
• Satellite/aerial views of land, for example to
determine how much land is being used for different
purposes, or to investigate the suitability of different regions
for different crops,
• inspection of fruit and vegetables distinguishing
good and fresh produce from old.
3. Industry
• Automatic inspection of items on a production line,
• inspection of paper samples.
4. Law enforcement
• Fingerprint analysis,
• sharpening or de-blurring of speed-camera images.
II. LITERATURE SURVEY
Automated grading has the potential to improve the
efficiency of diabetic retinopathy screening services. While
disease/no disease grading can be performed using only
micro aneurysm detection and image-quality assessment,
automated recognition of other types of lesions may be
advantageous. This study investigated whether inclusion of
automated recognition of exudates and hemorrhages
improves the detection of observable/referable diabetic
retinopathy. Automated detection of exudates and
hemorrhages improved the detection of observable/referable
retinopathy.[1]
To develop a technique to detect micro aneurysms
automatically in 50 degrees digital red-free Fundus
photographs and evaluates its performance as a tool for
screening diabetic patients for retinopathy. Candidate micro
aneurysms are extracted, after the image has been modified
to remove variations in background intensity, by algorithms
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
125 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
that enhance small round features. Each micro aneurysm
candidate is then classified according to its intensity and size
by the application of a set of rules derived from a training
set of 102 images. An automated technique was developed
to detect retinopathy in digital red-free Fundus images that
can form part of a diabetic retinopathy screening
programme. It is believed that it can perform a useful role in
this context identifying images worthy of closer inspection
or eliminating 50% or more of the screening population who
have no retinopathy.[2]
The robust detection of red lesions in digital color Fundus
photographs is a critical step in the development of
automated screening systems for diabetic retinopathy. In this
paper, a novel red lesion detection method is presented
based on a hybrid approach, combining prior works by
Spencer et al. (1996) and Frame et al. (1998) with two
important new contributions. The first contribution is a new
red lesion candidate detection system based on pixel
classification. Using this technique, vasculature and red
lesions are separated from the background of the image.
After removal of the connected vasculature the remaining
objects are considered possible red lesions. Second, an
extensive number of new features are added to those
proposed by Spencer-Frame.
III. PROBLEM FORMULATION AND
OBJECTIVE
Images are considered as one of the most important medium
of conveying information. The main idea of the image
segmentation is to group pixels in homogeneous regions and
the usual approach to do this is by „common feature.
Features can be represented by the space of color, texture
and gray levels, each exploring similarities between pixels
of a region.
Segmentation Algorithms
Image segmentation is the first step in image analysis and
pattern recognition. It is a critical and essential component
of image analysis system, is one of the most difficult tasks
in image processing, and determines the quality of the final
result of analysis. Image segmentation is the process of
dividing an image into different regions such that each
region is homogeneous. A.
K-means Clustering
K-Means algorithm is an unsupervised clustering algorithm
that classifies the input data points into multiple classes
based on their inherent distance from each other.[3]
Finally, this algorithm aims at minimizing an objective
function, in this case a squared error function [4].
Diabetic Retinopathy (DR) is one of the leading
causes of blindness in the industrialized world.
Early detection is the key in providing effective
treatment.
However, the current number of trained eye care
specialists is inadequate to screen the increasing
number of diabetic patients.
In recent years, automated and semi-automated
systems to detect DR with color fundus images
have been developed with encouraging, but not
fully satisfactory results.
IV. DESIGN METHODOLOGY
The proposed work is planned to be carried out in
the following manner
Fig 4: Working model of proposed system
Phase 1
Image Pre-processing
Images are enhanced by sharpening and removing
unwanted outliers.
Fig 5: Image processing
Phase 2
Segmentation
Image will be segmented to fetch out the image edges
and then detected all required parameters
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
126 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
Fig 6: Segmentation
Phase 3
Recognition and Classification
Ones the image is segmented it can be tested to
recognize it first and then classify it into original or
fraud image using SVM algorithm.
Fig. 7 SVM Algorithm
In the proposed work, we will develop a system to
detect fraud coins and notes currency for Indian Notes.
Clustering will be done using k-means algorithm.
Recognition and classification will be done using SVM
algorithm.
K-means Algorithm
k-means clustering is a method of vector quantization,
originally from signal processing, that is popular for
cluster analysis in data mining. k-means clustering aims
to partition n observations into k clusters in which each
observation belongs to the cluster with the nearest
mean, serving as a prototype of the cluster.
SVM Algorithm
In machine learning, support vector machines (SVMs,
also support vector networks) are supervised learning
models with associated learning algorithms that analyze
data used for classification and regression analysis.
V. DESIGN METHODOLOGY
Snapshots :
Fig.8 Front UI of the system.
In this form we have provided user with different GUI
components for accessing the proposed system.In above
form we have provided six buttons with four plots.
Fig.9 Uploading original image.
By clicking on input image we browse different fundus
images from the dataset.As you can see the image is then
plotted in of the respective slots provided in top left
corner.For capturing the image we have used imread
function from MATLAB and then we have used plot
function to display it on the screen.
Fig.10 Preprocessing of original image
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
127 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
By clicking on preprocessing the original image is converted
into gray scale and then Red Green and Blue channel are
extracted. The image is then displayed on the form using
implot function.
Fig.11 Feature Extraction
In feature extraction we use feature extraction algorithm for
extracting all the parameter from the processed image. This
parameter is consist of brightness, entropy, darkness,
gamma, and contrast.
Fig.12 Detection of retinopathy disease.
For detecting the disease in retina classification algorithm is
used to comapare the input image with the dataset images.
The algorithm compares all the features of input image
using cofusion matrix.
Fig.13 Detected part of hemorrhages
In this figure the detected part of hemorrhages is shown
using K-means theory.
Fig.14 Analysis of result.
This figure shows the percentage of data match in input
image and dataset. All the parameters are shown in bottom
right plot and rate of matching is also displayed.
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
128 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
Microanerysm:
Fig.15 Front UI of the system.
In this form we have provided user with different GUI
components for accessing the proposed system.In above
form we have provided six buttons with four plots.
Fig.16 Uploading original image.
By clicking on input image we browse different fundus
images from the dataset.As you can see the image is then
plotted in of the respective slots provided in top left
corner.For capturing the image we have used imread
function from MATLAB and then we have used plot
function to display it on the screen.
Fig.17 Preprocessing of original image
By clicking on preprocessing the original image is converted
into gray scale and then Red Green and Blue channel are
extracted. The image is then displayed on the form using
implot function.
Fig.18 Preprocessing of original image
Fig.19 Detection of retinopathy disease.
For detecting the disease in retina classification algorithm is
used to comapare the input image with the dataset images.
The algorithm compares all the features of input image
using cofusion matrix.
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
129 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
Fig.20 Detected part of Microneurysm
In this figure the detected part of hemorrhages is shown
using K-means theory.
VI. RESULTS AND DISCUSSION
Performance parameter for evaluation:
In pattern recognition, information retrieval and binary
classification, precision (also called positive predictive
value) is the fraction of relevant instances among the
retrieved instances, while recall (also known as sensitivity)
is the fraction of relevant instances that have been retrieved
over the total amount of relevant instances. Both precision
and recall are therefore based on an understanding and
measure of relevance.
Precision can be seen as a measure of exactness or quality,
whereas recall is a measure of completeness or quantity.
Table 1. Performance Evaluation
Test
Result
Present
Absent
Positive True Positive (TP) False Positive (FP)
Negative False Negative
(FN)
True Negative
(TN)
𝑇𝑃
𝑆𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 (Precision)=
________________________________________________
______
𝑇𝑃+𝐹P
𝑇P
𝑆𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑖𝑡𝑦
(Recall)=_________________________________________
______________
𝑇P+𝐹𝑃+TN+FN
Table 2 shows the sensitivity and specificity of various
grades of images from the dataset images and the results are
also compared with the expert opinion. Out of 32 sever
images all 32 is detected as sever, out of 18 moderate
images all 18 are detected as moderate, out of 24 mild
images, 3 images were detected as healthy and out of 26
healthy images 3 images are detected as mild. Figure 8
shows the graph of Sensitivity and Specificity of Grade 0,
Grade1, Grade 2 and Grade 3 images from the dataset.
The average sensitivity and specificity for the proposed
ADDR is 94.44% and 87.5% respectively. Table 5 shows
the comparison results with the other papers.
Table 2. Grading of fundus image analysis
system on the Dataset
DR
Grade
Number
of
images
Sensitivity Specificity
Grade 0 26 89.65% 100%
Grade 1 24 100% 88.88%
Grade 2 18 100% 100%
Grade 3 32 100% 100%
VII. CONCLUSION AND FUTURE SCOPE
Conclusion
Prolonged diabetes leads to DR, where the retina is damaged
due to fluid leaking from the blood vessels. Usually, the
stage of DR is judged based on blood vessels, exudes,
hemorrhages, microaneurysms and texture. In this paper, we
have discussed different methods for features extraction and
automatic DR stage detection. An ophthalmologist uses an
ophthalmoscope to visualize the blood vessels and his or her
brain to detect the DR stages. Recently digital imaging
became available as a tool for DR screening.
Future Scope
A main weakness of the algorithm arises from the fact that
the algorithm depends on vessel detection. This indicates the
further necessity of improving the robustness of this task.
Hemorrhages detection could be also added to the system in
order to increase its ability to verify the degree of diabetic
retinopathy.
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 5 Issue: 6 123 – 130
_______________________________________________________________________________________________
130 IJFRCSCE | June 2019, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
REFERENCES
[1] S. Wild, G. Roglic, A Green, “Global prevalence of diabetes:
estimates for the year 2000 and projections for 2030,”
Diabetes Care, 27, pp.l047- 1053, 2004.
[2] S. R. Nirmala, M. K. Nath, and S. Dandapat, “Retinal Image
Analysis: A Review,” International Journal of Computer &
Communication Technology (IJCCT), vol-2, pp. 11-15,
2011.
[3] National Eye Institute, National Institutes of Health,
“Diabetic Retinopathy: What you should know,” Booklet,
NIH Publication, no: 06-2171, 2003.
[4] A. D. Fleming, K. A. Goatman, and J. A. Olson, “The role of
haemorrhage and exudate detection in automated grading of
diabetic retinopathy,” British Journal of Ophthalmology, vol
94, no. 6, pp. 706- 711, 2010.
[5] A. D. Fleming, S. Philip, K. A. Goatman, J. A. Olson, and P.
F. Sharp, “Automated microaneurysm detection using local
contrast normalization and local vessel detection,” IEEE
Trans. Med. Imag., vol. 25, no. 9, pp. 1223–1232, Sep. 2006.
[6] M.Niemeijer,B. V. Ginneken, J. Staal, M. S. A. Suttorp-
Schulten, and M. D. Abramoff, “Automatic detection of red
lesions in digital color fundus photographs,” IEEE Trans.
Med. Imag., vol. 24, no. 5, pp. 584– 592, May 2005.
[7] M. J. Cree, J. A. Olsoni, K. C. McHardyt, J. V. Forresters
and P. F. Sharp, “Automated microaneurysms detection,”
IEEE conference, pp. 699-702, 1996
[8] A. Mizutani, C. Muramatsu, Y. Hatanaka, S. Suemori, T.
Hara, and H. Fujita, “Automated microaneurysm detection
method based on double ring filter in retinal fundus images,”
in Proc. SPIE, Med. Imag, Imag,Computer-Aided Diagnosis,
vol. 7260, 2009.
[9] K. Ram, G. D. Joshi and J. Sivaswamy, “A successive clutter
rejectionbased-approach for early detection of diabetic
retinopathy,” IEEE Transaction on Biomediacl Engineering,
vol. 58, no. 3, pp. 664-673, Mar. 2011
[10] M. Esmaeili, H. Rabbani, A. M. Dehnavi, and A. Dehghani,
“A new curvlet transform based method for extraction of red
lesions in digital color retinal images,” in Proc. ICIP,
pp.4093-4096, 2010.
[11] T. Spencer, J. Olson, K. McHardy, P. Sharp, and J. Forrester,
“An image processing strategy for the segmentation and
quantification in fluorescein angiograms of the ocular
fundus,” Comput. Biomed. Res., vol.9, pp. 284–302, 1996.