IRJET- A Novel Algorithm for Detection of Papilledema in Luminosity and Contrast Enhanced Retinal Images

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 04 | Apr 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 4464
A NOVEL ALGORITHM FOR DETECTION OF PAPILLEDEMA IN
LUMINOSITY AND CONTRAST ENHANCED RETINAL IMAGES
ANEESA P A1, SAJITHA S2
1M.Tech.student, Applied Electronics and Communication system, Nehru College of engineering and research
centre, Kerala, India
2Professor, Electronics and Communication department, Nehru College of engineering and research centre, Kerala,
India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Papilledema is a disease caused by the
intracranial pressure due to any reason and it will results in
swelling of optic disc. Enhanced color retinal images are
taken to detect the presence of this disease. Normal images
with poor quality cannot be used for diagnosis .Quality
degradation may be due to uneven illumination, image
blurring, low contrast etc. luminosity and contrast
enhancement is applied on the poor quality retinal images
.R, G and B channels in the color retinal image is enhanced
by the luminous gain matrix which is the obtained by
gamma correction of the value channel in HSV color space.
Contrast is enhanced by CLAHE (contrast limited adaptive
histogram equalization) in the luminosity channel of L*a*b*
color space. In this enhanced image, optic disc is located.
Papilledema is indicated by the swelling of optic disc.So
optic disc (OD) is needed to be located at first .OD is the
brightest region in retina. It is the exit point of retinal nerve
fibers from eye and the entrance and exit point for retinal
blood vessels. OD is located using swarm intelligence. When
the number of pixels in the located OD is greater than a
threshold value, the disease is said to be detected otherwise
not. It is very effective as well as time saving work.
Key Words: Gamma correction, L*a*b*colour space,
optic disc, swarm intelligence, Luminosity
1. INTRODUCTION
There are many diseases associated with the retina.
Papilledema is one of such diseases. It is due to the
intracranial pressure. It will result in swelling of the optic
disc .The efficient means of detecting papilledema is valid
in nowadays where people are common to many eye
diseases. In order to detect papilledema the retinal images
from the patients must be analyzed. These images are
having poor quality due to uneven illumination, image
blurring, low contrast etc. So these images must be
enhanced. There are many enhancement methods from
that enhancement based on luminosity and contrast
adjustment is used here.Enhancement is carried out in
color retinal images. By performing gamma correction in
the value channel of the HSV (hue, saturation, and value)
color space will results in a luminance gain matrix and it is
used to enhance the R,G and B (red green and blue)
channels. Contrast is then enhanced in the luminosity
channel of L*a*b* color space by CLAHE (contrast limited
adaptive histogram equalization). In recent years many
enhancement methods are proposed including image
luminosity and contrast normalization techniques, A multi
scale method based on the contour let transform, CLAHE
(contrast limited adaptive histogram equalization),
Retinex-based enhancement algorithm etc. These methods
are used to provide enhancement for retinal blood vessels
by exhibiting greater contrast between blood vessels and
the retinal background in both gray scale and color retinal
images. Here the green channels of the color retinal image
displays a high contrast between the vessels and the
background these enhanced retinal images lose color
information or other important image features. So the
analysis of the ophthalmologists gets degraded. The retinal
diseases are analyzed using the measurement of optic disc
size and the optic cup size. By separating the optic disc, it
will be helpful for analyzing the nerve vessels. By
analyzing these nerve vessels various retinal diseases such
as glaucoma, cataract, and cardiac disease can be detected
[1][2]. By performing enhancement of retinal images
based on luminosity and contrast enhancement we can
distinguish various anatomical structures like the macula,
optic disc (OD), optic cup and blood vessels and also it can
provide increased quality without color distortion and
over enhancement. To avoid the common problem of color
distortion, all the processes are performed on the
luminosity channel. Now, coming to the detection of
papilledema which is optic disc swelling. The optic disc is
needed to be located at first. The idea of swarm
intelligence is used for the segmentation of the optic disc.
When the number of pixels in the located optic disc is
greater than the threshold value, papilledema is said to be
detected otherwise not.
2. LITERATURE SURVEY
There are various methods and algorithms which are
implemented in fundus images. By calculating the optic
disc the disease of the fundus images can be identified.
Huiqi Li et al. [3] have implemented Principal Component
Analysis(PCA) to find out the lesions in the retinal images.
Optic disc layer can be found by the modified active shape

model algorithm and this technique has disadvantaged one
due to the slow processing. Hoover [4] implemented the
geometric analysis of optic disc and the retinal blood
vessels to separate the disc location in the retina. The
method used in the above paper is the fuzzy logic to find
the optic disc and optic cup from the fundus image
3. PROPOSED WORK
The proposed technique consists of following block
diagram. The block diagram represented in the figure 1
Fig- 1: Block diagram of proposed work
3.1 IMAGE ACQUISITION
The images are obtained from nearby hospital. From that
a number of images which are normal as well as diseased
are selected. The input image of the retina is shown in the
figure.2
Fig-2: Input image
3.2 PRE-PROCESSING
A number of Pre-processing steps should be followed
on the retinal images before Optic Disc localization. These
are given by
• Resizing
• Grey scale conversion
• Median filtering
• Adaptive histogram equalization
• Background subtraction
• Median filtering
• Mean filtering
3.3 IMAGE ENHANCEMENT
ENHANCEMENT OF COLOUR RETINAL IMAGE
It consists of two steps: Luminosity enhancement and
contrast enhancement
LUMINOSITY ENHANCEMENT
Luminance can degrade visual quality. So it is essential
to enhance luminous effect. Color retinal images are split in
to R, G, B channels then it is converted in to HSV (hue,
saturation and value) color space. The luminosity channel
is used for all processing, performing gamma correction on
luminosity channel will results in luminance gain matrix.
This luminance gain matrix is used to enhance RGB
channels. Gamma correction is an image processing
method, here the luminance is transformed nonlinearly,
and the transformation curve is
w =u^r
Where u has the values in the range (0, 1) and it is the
normalized fixed value of the luminosity channel. W is the
normalized output and ‘r’ is a constant
CONTRAST ENHANCEMENT
CLAHE method is used to enhance contrast of the Color
retinal images. In this method retinal image is divided into
small regions known as tiles. The local contrast is
enhanced in such a way that the histogram on each tile is
equalized here. The luminosity enhanced R, G, B channel is
converted in to L*a*b* color space. Then CLAHE is applied
on the luminosity channel and again color space
conversion is carried out. That is converted in to RGB.
Then channel is merged and the enhanced color retinal
image is obtained. The enhanced image is shown in the
figure 3
Fig- 3: Enhanced image
3.4 IMAGE SEGMENTATION
OPTIC DISC LOCALISATION
Optic Disc is the brightest region in the retina, and it is
the exit point of retinal fibers. Papilledema is affected to
the Optic Disc. The location of OD is very important

.
3.5 FEATURE EXTRACTION
PARTICLE SWARM OPTIMIZATION
PSO is a computational method that optimize a problem
by iteratively trying to improve a candidate solution with
regard to a given measures of quality. In the location of
optic disc there is a group of particles. Optic disc is the
brightest region so the intensity of the pixels will be very
high. PSO algorithm is used to locate optic disc. In particle
swarm optimization the swarm is modeled by particles
that can move along the multidimensional space each
particle has its own velocity and position. The particles
move towards the best position each particles aim is the
global best position which is the desired position. This
global best is updated when a new best position is known.
Every particle tends to follow another particle which is
nearest to the global best. Local best is the best position of
each particle. Every particle communicates with each
other .Neighborhood best is the best position obtained by
communicating with another particle in the swarm.
The constriction coefficients are
Phi 1=20.5
Phi 2=20.5
Phi 1=20.5=Phi 1=20.5+Phi 1=20.5
Chi=2/ (phi-2+sqv + (phi^2-4*phi)
W=chi inertia weight
Wdamp = 1 Inertia weight dumping ratio
C1=chi*phi1 personal learning coefficient
C2= chi*phi2 global learning coefficient, the population size
is row *Colum.
After locating the optic disc, a circle is drawn on the
location of the optic disc to identity it clearly
3.6 CLASSIFICATION
PAPILLEDEMA DETECTION
After locating the optic disc it is converted into binary
form. In that optic disc having a round shape with white
colour and background is black. A predefined threshold is
set. Whenever the number of pixels in the white area is
greater than this threshold the disease is said to be
detected otherwise not.
4. RESULTS
Thus, the final optic disc swelling is detected from the
diseased patient. The accuracy of the proposed work tends
to be 92% and the sensitivity is around 40% and
specificity is around 92%.The detected swelling is shown
in the figure 4.
Fig-4: OD Swelling detected
5. CONCLUSIONS
Thus, the proposed method can detect the various
patients’ retinal disease at the less time period. In future
by increasing the training and testing periods more
number of diseases can be found at the same time using
retinal images.
REFERENCES
[1] Juan Xu, Hiroshi Ishikawa, Gadi Wollstein, Richard A.
Bilonick, Kyung R. Sung,Larry Kagemann, Kelly A.
Townsend, and Joel S. Schuman, “Automated Assessment
of Optic Nerve Head on Stereo Disc Photographs”, Invest
Ophthalmol Vis Sci. Jun 2008; 49(6), pp. 2512– 2517,
2008.
[2] Thitiporn Chanwimaluang and Guoliang Fan, “An
efficient algorithm for extraction of anatomical structures
in retinal images”, Proc. of International Conference on
Image Processing, Vol. 1, pp. 1093– 1096,2003
[3] Huiqi Li, Opas Chutatape “Boundary detection of optic
disc by a modified ASM method”, The Journal of the
Pattern Recognition Society, Vol. 36, pp. 2093-2104,2003
[4] Hoover, A., Goldbaum, M., “Locating the optic nerve in a
retinal image using the fuzzy convergence of the blood
vessels”, IEEE Trans. on Medical Imaging 22(8), 951–958
(2003)
[5] Mendels, F., Heneghan, C., Thiran, J.P., “Identification of
the optic disc boundary in retinal images using active
contours”, In: Proc. IMVIP conference, pp. 103–115 (1999)
[6] Osareh, A., Mirmehd, M., Thomas, B., Markham, R.,
“Comparison of color spaces for optic disc localisation in
retinal images”, In: 16th International Conference on
Pattern Recognition, vol. 1, pp. 743–746 (2002)
[7] Li, H., Chutatape, O., “Automated feature extraction in
color retinal images by a model based approach”, IEEE
Trans. on Biomedical Engineering 51(2), 246–254 (2004)
[8] Li, H., Chutatape, O., “ Boundary detection of optic disk
by a modified ASM method”, Pattern Recognition 36(9),
2093–2104 (2003)
[9] Jun Cheng, Jiang Liu , Wong,D.W.K. , Fengshou Yin ,
“Automatic optic disc segmentation with peripapillary
atrophy elimination”, Proc. EMBC, Annual Conference,
IEEE, Aug. 30 - Sept. 3 2011, pp. 6224 – 6227.
[10] Handayani Tjandrasa, Ari Wijayanti, Nanik Suciati
,“Optic Nerve Head Segmentation Using Hough Transform

and Active Contours”, TELKOMNIKA, Vol.10, No.3, July
2012, pp.531~536.
[11] Rafael C Gonzalez, Richard E Woods, Steven L Eddins,
Digital Image Processing, Prentice Hall Publications, 2009.
[12] Rafael C Gonzalez, Richard E Woods, Steven L Eddins.,
Digital Image Processing Using Matlab, Prentice Hall
Publications, 2009.
[13] Yamille del Valle, Ganesh Kumar Venayagamoorthy,
Salman Mohagheghi, Jean-Carlos Hernandez, Ronald G.
Harley, “Particle Swarm Optimization: Basic
Concepts,Variants and Applications in Power Systems”,
IEEE Transactions On Evolutionary Computation, Vol. 12,
NO. 2, pp. 171- 195 APRIL 2008
[14] Chih-Chin Lai, “A Novel Image Segmentation
Approach Based on Particle Swarm Optimisation”, IEICE
Trans.Fundamentals, Vol. E89 A, No.1 pp. 324-327,
January 2006

IRJET- A Novel Algorithm for Detection of Papilledema in Luminosity and Contrast Enhanced Retinal Images

More Related Content

What's hot (19)

Similar to IRJET- A Novel Algorithm for Detection of Papilledema in Luminosity and Contrast Enhanced Retinal Images (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- A Novel Algorithm for Detection of Papilledema in Luminosity and Contrast Enhanced Retinal Images