International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
Design Strategies for Classification of
Abnormalities in Retinal Images Using ANFIS
Manisha P. Waghmare 1, Dr. S. D. Chede2, Prof. S. M. Sakhare3
1
Student M. Tech, Department of Electronics, Suresh Deshmukh College of Engineering Selukate Wardha
2
Principal, Om College of Engineering Wardha
3
Assistant Professor, Department of ECE, Suresh Deshmukh College of Engineering Selukate Wardha
Abstract
The human eye is the organ which reacts to light and gives us the sense of sight. The eye emits or reflects the light to interpret the
colors, shapes, and dimensions of objects in the world. Retina plays a major role in the human vision system. Many important eye
diseases manifest themselves in the retina. While some other anatomical structures contribute to the process of vision in eye, this
review focuses on retinal image and image analysis. The causes are Retinopathy of Prematurity, Diabetic Retinopathy,
Hypertensive Retinopathy, and Obstruction of arterial Circulation, Sickle Cell Retinopathy and Obstruction of the venous
Circulation. This work aims to detect some abnormalities in the retinal image and to classify those using ANFIS. The
methodology is used to Preprocessing, Candidate Extraction, Feature Extraction, Classification and Performance analysis. This
paper gives the idea about the Preprocessing. The results are obtained by using MATLAB software.
Keywords: Retina, Retina Diseases, ANFIS, Biomedical image processing, Preprocessing.
1. INTRODUCTION
The human eye is the organ which reacts to light and gives us the sense of sight. Rod and cone cells in the retina allow
conscious light perception and vision including color differentiation. The human eye can distinguish about 10 million
colors. Many important eye diseases manifest themselves in the retina. While a number of other anatomical structures
contribute to the process of vision in eye, this review focuses on retinal image and image analysis [1]. Our aim is to
classify the diseases of retina like Hypertensive Retinopathy, Diabetic Retinopathy, Retinopathy of Prematurity, Sickle
Cell Retinopathy, Obstruction of arterial Circulation, and Obstruction of the venous Circulation using one of the artificial
intelligence technique.
Hypertensive Retinopathy
This may occur under four circumstances. In simple hypertension without sclerosis, as seen in young patients, the retinal
signs are few: a generalized constriction of the arterioles which appear to be pale and unduly straight with acute-angled
branching, additional superficial, flame-shaped hemorrhages and cotton-wool spots may occur and hard exudates are
absent [9].
Diabetic Retinopathy
Diabetic retinopathy is the most common diabetic eye disease and a leading cause of blindness. It is caused due to changes
in the blood vessels of retina. In some people blood vessels may leak fluid on the retinal surface due to diabetic
retinopathy. In other people, new blood vessels grow on the surface of the retina. The retina is the Light-sensitive part at
the back of the eye. A long term diabetic retinopathy can get worse and cause loss of vision. Diabetic retinopathy affects
both the eyes [9].
Retinopathy of Prematurity
The retinal manifestations of this disease are generally noted some weeks after birth in premature infants who have been
given high concentration of oxygen. The earliest signs are dilatation of the retinal veins and the appearance of hazy white
patches in the periphery of the retina, which soon show an indefinite proliferation into the vitreous. This is due to the
formation of new vessels in the retina itself, which bud into the vitreous. Their appearance is followed by the development
of fibrous tissue, which eventually proliferates to form a continuous mass behind the lens, appearing as a type of
Pseudoglioma [9].
Sickle Cell Retinopathy
Sickle cell hemoglobin is abnormal hemoglobin found mainly but not exclusively in people of African origin. When it is
deoxygenated it becomes insoluble and distorts the normally discoid red cell into a characteristic sickle shape. Such
suckled red cells tend to obstruct capillaries and this leads to infraction, particularly in the periphery of the retina [9].
Obstruction of Arterial Circulation
Central Retinal Artery Occlusion (CARO) is nearly always at the lamina cribrosa, where the vessels normally become
slightly narrowed. Such as accident causes sudden and complete retinal ischemia and this tissue rapidly die. The eye
becomes suddenly blind [9].
Obstruction of the Venous Circulation
Volume 3, Issue 3, March 2014
Page 388
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
In Central Retinal Vein Occlusion (CRVO), all the veins of the retina become enormously engorged with blood and
extremely tortuous, and the retina is covered with hemorrhages. In many cases bunches of tortuous new vessels are
formed upon the optic disc; in others a collateral circulation is effected by similarity tortuous new vessels in the retina [9].
In this paper, we detect this abnormality by using preprocessing method.
2. PROBLEM DEFINITION
Huge work has been done on retinal images for detection of abnormalities or diseases such as Hypertensive Retinopathy,
Diabetic Retinopathy, Retinopathy of Prematurity, Sickle Cell Retinopathy, Obstruction of arterial Circulation, and
Obstruction of the venous Circulation. Most of the work is done on preprocessing on images or detection of single
diseases. None of the existing work has targeted for a classification of more than one retinal disease. Hence the
sophisticated technique for classification of retinal diseases/abnormalities is to be investigated.
Through this work we are going to propose a solution for a classification of disease from retinal images like Hypertensive
Retinopathy, Diabetic Retinopathy, Retinopathy of Prematurity, Sickle Cell Retinopathy, Obstruction of arterial
Circulation, and Obstruction of the venous Circulation using one of the artificial intelligence technique.
3. PREPROCESSING
Preprocessing is a method which is used to remove the noise and enhance the image quality. Poor quality image is due to
patient movement, poor focus, bad positioning, reflections and inadequate illumination. Preprocessing improve the
automated abnormality detection. The retinal images collected from the data base are color images. Then the color images
are converted to gray scale images. Because the retinal abnormalities have better visualization in the gray scale when
compared to others. Then the preprocessing techniques are applied to the gray scale image.
4. EXPERIMENTAL RESULT
The experimental results of preprocessing methods are Intensity Equalization and Histogram Equalization.
Gray Scale Image: Grayscale digital image is an image in which the value of each pixel is a single sample that is it
carries only intensity information. Gray scale gives the better visualization as compared to the other. Gray scale images as
shown in fig (b) & Histogram plot for gray scale as shown in fig (e).
Intensity Equalization: The contrast in a low contrast grayscale image by remapping the data values to fill the entire
intensity range [0,255].Intensity equalization images as shown in fig (c).
Histogram Equalization: Histogram equalization is the technique by which the dynamic range of the histogram of an
image is increased. It improves contrast of an image. Histogram equalization images as shown in fig (d) and Histogram
plot for histogram equalization as shown in fig (f).
Normal Retina
Fig (a): Input Image (RGB)
Fig (b): Output Image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
(Gray scale)
Fig (e): Histogram plot for output image
Volume 3, Issue 3, March 2014
Fig (f): Histogram plot for histogram equalization
Page 389
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
Diabetic Retinopathy
Fig (a): Input Image (RGB)
Fig (b): Output Image
(Gray scale)
Fig (e): Histogram plot for output image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
Fig (f): Histogram plot for histogram equalization
Hypertensive Retinopathy
Fig (a): Input Image (RGB)
Fig (b): Output Image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
(Gray scale)
Fig (e): Histogram plot for output image
Volume 3, Issue 3, March 2014
Fig (f): Histogram plot for histogram equalization
Page 390
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
Retinopathy of Prematurity
Fig (a): Input Image (RGB)
Fig (b): Output Image
(Gray scale)
Fig (e): Histogram plot for utput image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
Fig (f): Histogram plot for histogram equalization
Sickle Cell Retinopathy
Fig (a): Input Image (RGB)
Fig (b): Output Image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
(Gray scale)
Fig (e): Histogram plot for output image
Volume 3, Issue 3, March 2014
Fig (f): Histogram plot for histogram equalization
Page 391
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
Obstruction of Arterial Circulation
Fig (a): Input Image (RGB)
Fig (b): Output Image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
(Gray scale)
Fig (e): Histogram plot for output image
Fig (f): Histogram plot for histogram equalization
Obstruction of Venous Circulation
Fig (a): Input Image (RGB)
Fig (b): Output Image
Fig (c): Intensity Equalization
Fig (d): Histogram Equalization
(Gray scale)
Fig (e): Histogram plot for output image
Fig (f): Histogram plot for histogram equalization
5. CONCLUSION
The retinal abnormalities have better visualization in the gray scale when compared to others. Using preprocessing
method on gray scale image, noise is removed from the image. Image quality get enhanced and improved with the help
Volume 3, Issue 3, March 2014
Page 392
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 3, March 2014
ISSN 2319 - 4847
of preprocessing method. Now the next step in classification of abnormalities in retinal images using ANFIS is to perform
the candidate extraction so as to extract the main infected area. A defected retina is compared with normal retina.
In a Diabetic Retinopathy, changes in the blood vessels of retina. In a Hypertensive Retinopathy, flame-shaped
hemorrhages and cotton-wool spots may occur and hard exudates are absent. In a Retinopathy of Prematurity, dilatation
of the retinal veins and the appearance of hazy white patches in the periphery of the retina. In a Sickle Cell Retinopathy,
this leads to infraction, particularly in the periphery of the retina. In a Obstruction of Arterial Circulation, vessels
normally become slightly broaden. In a Obstruction of the Venous Circulation, bunches of tortuous new vessels are
formed upon the optic disc.
References
[1] T.Yamuna and S.Maheswari “Detection of Abnormalities in Retinal Images”, IEEE International Conference on
Emerging Trends in Computing, Communication and Nanotechnology (ICECCN 2013).
[2] G.S. Annie Grace Vimala and S. Kaja Mohideen “Automatic Detection of Optic Disk and Exudate from Retinal
Images Using Clustering Algorithm” IEEE Proceedings of7'h International Conference on Intelligent Systems and
Control (ISCO 2013)
[3] Sina Hooshyar, Rasoul Khayati “Retina Vessel Detection Using Fuzzy Ant Colony Algorithm” 2010 IEEE Canadian
Conference Computer and Robot Vision.
[4] Alireza Osareh, Bita Shadgar, and Richard Markham “A Computational-Intelligence-Based Approach for Detection
of Exudates in Diabetic Retinopathy Images”, IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN
BIOMEDICINE, VOL. 13, NO. 4, JULY 2009.
[5] M. I. Iqbal1, A. M. Aibinu2, M. Nilsson1, I. B. Tijani2, and M. J. E. Salami “Detection of Vascular Intersection in
Retina Fundus Image Using Modified Cross Point Number and Neural Network Technique”, Proceedings of the
International Conference on Computer and Communication Engineering 2008 IEEE.
[6] Ms. Rupa V. Lichode, Prof. P. S. Kulkarni “Automatic Diagnosis of Diabetic Retinopathy by Hybrid Multilayer
Feed Forward Neural Network”, International Journal of Science, Engineering and Technology Research (IJSETR)
Volume 2,Issue 9, September 2013.
[7] Neera Singh, Ramesh Chandra Tripathi “Automated Early Detection of Diabetic Retinopathy Using Image Analysis
Techniques”, International Journal of Computer Applications (0975 – 8887) Volume 8– No.2, October 2010.
[8] Asha Gowda Karegowda, Asfiya Nasiha and M.A.Jayaram “Exudates Detection in Retinal Images using Back
Propagation Neural Network”, International Journal of Computer Applications (0975 – 8887) Volume 25– No.3, July
2011.
[9] Parson’s text book of Ophthalmology
AUTHOR
Miss Manisha P. Waghmare received the bachelor degree B.E in Electronics Engg. From R.T.M.N.U Nagpur
university in the year 2011.Currently she is working as Research scholer for pursuing M.Tech in Electronics
(Communication Engineering) at SDCE,Selukate,Wardha,Maharashtra,India
Dr.Santosh D. Chede received the bachelor degree B.E in Industrial Electronics from Amravati university in
the year 1990.Also he has received his master’s degree M.E in electronics Engg. From Amravati university in
the year 2000.In year 2010 he received his Doctor of Philosophy from VNIT,Nagpur.Currently he is working
as Principal, Om college of Engg,Inzapur, Wardha, Maharashtra,India.
Prof. Shailesh M. Sakhare received the bachelor of Engg. Degree B.E in Electronics Engg. From R.T.M
Nagpur university in 2008.Also he received his master’s degree master of Technology in Electronics Engg in
2012 from GHRCE,Nagpur,An Autonomous Institute . Currently he is working as Asst.Professor in
Electronics & communication Dept. at SDCE,Selukate,Wardha,Maharashtra,India.
Volume 3, Issue 3, March 2014
Page 393