ifferent Techniques for Cataract Detection

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 02 | Feb -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 366
Different Techniques for Cataract Detection
Anjali K1, Bhavya K Bharathan2, Hanan Hussain3, Nirmala P S4, Swathy M5
1PG Scholar, Dept. of Computer Science &Engineering, Vidya Academy of Science & Technology, Kerala, India
2 PG Scholar, Dept. of Computer Science &Engineering, Vidya Academy of Science & Technology, Kerala, India
3 PG Scholar, Dept. of Computer Science &Engineering, Vidya Academy of Science & Technology, Kerala, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Cataract is an eye disease that caused by
opacity of lens. This will leads to complete loss of
vision. Mostly cataracts are affected to aged peoples.
In this modern era, so many techniques are used for
detecting cataract. In this paper, various techniques
for the diagnosis of cataract are focused.
Key Words: Cataract, Feature extraction, Thresholding,
Classification
1.INTRODUCTION
Cataracts are very general in older people.The main
causes of cataracts are diabetes, optic nerve damage, and
macular degeneration and it can occur in either or both
eyes.
It cannot spread from one eye to the other. It is mainly
affected to lens of retina. The lens is a clear part of the eye
that helps to focus light, or an image, on the retina. In a
normal eye, light passes through the transparent lens to
the retina. Range of affected people is depicted in figure 1.
Fig -1: Range of cataract affected people
1.1 Types of Cataract
Cataracts classified based on cause:-
 Secondary cataract: Cataracts also can develop in
people who have other health problems, such as
diabetes.
 Traumatic cataract: It can develop after an eye injury,
sometimes years later.
 Congenital cataract: Some times cataracts develop in
childhood, often in both eyes. It may be so small that
they do not affect vision.
 Radiation cataract: It can develop after exposure to
some types of radiation.
Cataracts classified based on age:-
 Congenital and acquired
Cataracts classified based on location:-
 Cortical, nuclear, sub-capsular
Cataracts classified based on shape:-
 Dot-like, coronary, lamellar
Cataracts classified based on degree:-
 Immature, intumescent, mature, hypermature
1.2 Symptoms
The most common symptoms of a cataract are:
 Cloudy or blurry vision and poor night vision
 Glare headlights, sunlight or lamps
 Double vision or multiple images in one eye
 Frequent prescription changes in your eyeglasses
1.3 Causes
Most common causes for cataract are:
 Lifestyle, age, and diet
 Previous eye injury [1]
 Limit overexposure to sunlight
 High altitude can also contribute to cataracts
 Diabetes, obesity, or high blood pressure [2][3]
 Smoking or drinking too much alcohol[4]
Difference between normal eye and cataract eye are
shown in figure 2.

Fig -2: Normal eye and eye with cataract
Cataracts can be not easy to detect. The opacity of the
lens of eye may be obvious to many people; it may not be
noticeable to others.
2.CATARACT TESTS AND TREATMENTS
The existing tests and treatments for cataract detection
are shown below:
2.1 CATARACT TESTS
The amount of visual deterioration, representing the
asperity of the cataract, can be obtained by an eye
examination which may contain the following
investigations.
Refraction test: This test will finds whether glasses can
help to enhance the vision.
Visual acuity test: A visual awareness eye test is identical
to the routine eye test done throughout life by an
ophthalmologist. Both eyes are tested individually by
using a viewing device or an eye chart in order to
determine the ability to see letters of gradually reduced
sizes. Using this method, the doctor can understand what
extent the vision has been affected by the cataract. Visual
acumen is measurement of how well a person can see.
Contrast sensitivity testing: Test is similar to visual
acumen testing but it shows more definitely the reduced
image contrast caused by a cataract, as result of light glare
and scattering caused by the cataract. The capability to
distinguish between various shades of gray forms the
basis of this test since this ability may be blocked in the
existence of a cataract.
Color vision testing: Helps to detect obtained color vision
defects that can be seen in cataract patients.
Glare Testing: Vision may be changed based upon
different lighting conditions, such as at night and in
brilliant sunshine. These marks may be found out with
various types of lighting by having a patient read the chart
twice, once with and without bright lights.
Potential acuity testing: Test that can give an almost idea
about the vision following cataract removal and taken as
the eye’s vision power if there was no cataract.
Spectacular photographic microscopy: To take a
photograph of the endothelial layer of the cornea, a
specialized microscope is used. This is usually done
previous to cataract surgery to find out the health of the
endothelium, which is likely to affect the outcome of
surgery.
Retinal examination: Before this examination, the pupils
are dilated with eye drops so that the retina may be better
visualized. An ophthalmoscope or slit-lamp is used to look
for signs of cataract, as well as signs of macular
degeneration, glaucoma, and other problems related to the
optic nerves and retina which could be the cause of vision
impairment.
Slit-lamp examination: Done with a special microscope
known as the slit-lamp, which projects an intense, thin
beam of light into the eye to give an amplified three-
dimensional view of the interior of the eye. Manual
detection is able to examine in section the structures at the
front of the eye, including the iris, cornea and lens, as well
as the area between the cornea and iris and look for any
disorders.
Tonometry: Test may be done to calculate the pressure
within the eye, or intraocular pressure (IOP), by a special
instrument. Eye drops may be injected before doing the
test. Raised IOP may represent glaucoma.
2.2 CATARACT TREATMENTS
Treatment in the early stages may be simple, such as
the use of alternative lenses, eyeglasses, alternative lenses,
anti-glare sunglasses, aalternative lenses or just an
adjustment in environmental lighting.
Medication: Drug therapy cannot heal a cataract.
Mydriatic eye drops which dilate the pupils may help in
some cases for a short span of time by raising the amount
of light entering the eye. Ssometimes recommended for
young children who are waiting for cataract surgery so as
to avoid vision loss in the interim.
Surgery: It is only effective treatment option in the case of
more severe cataracts, especially when it affects daily
actions or when it is combined with other problems.
Removal of cataract surgery is safe and highly accurate in
enhancing vision. And is not damaged significantly then
surgery may not necessary once a cataract is detected.
During surgery, the cloudy lens is removed and is replaced
with an artificial lens. Whether there are cataracts in both
the eyes which require treatment, surgery is usually done
one at a time with an interval of four to eight weeks
between the two processes. Phacoemulsification is the
general process for cataract extraction, the cloudy lens is
destroyed up by a probe which emits ultrasonic vibrations
and the particles are then removed by suction. Other
surgical method is called extra-capsular cataract surgery,
where the cloudy lens is removed as a whole. Intra-

capsular cataract surgery is rarely done nowadays, where
the lens, along with its capsule, is removed. Surgery is
gradually done on an outpatient basis. The different
surgical procedures are discussed in detail under cataract
lens removal and replacement surgery.
3.CATARACT DETECTION TECHNIQUES
Rafat et.al [5], the technique of DLS or dynamic light
scattering is used for detection of cataract at molecular
level. However, the victory of this system in experimental
use depends upon the exact control of the dispersal
volume inside a patient's eye and particularly during
patient's replicate visits. This is significant because the
dispersal volume within the eye in a high-quality DLS set-
up is very less. A corneal analyzer was customized by
introducing a DLS fiber optic imaging probe within its
cone. Figure 3 is a schematic illustration of the optical
system.
Fig -3: Diagram of the optical system
The unmistakable data obtained in this study is
significant in planning a longitudinal study of anti-cataract
drug screening.
Nayak et.al [6], pre-processing is made to diminish the
contrast and to regulate the mean intensity. Intensities of
the three colour bands were altered to an intensity-hue
saturation representation [7]. It allows the intensity to be
processed without affecting the professed relative colour
values of the pixels. Mainly the features of the optical eye
image such as small ring area, big ring area, object
perimeter and edge pixel count are extracted. The inner
surface of the cornea images is more whitish relative to
that of normal and post-cataract images. It is the origin for
manipulating small ring area. Colour at the external
surface of the cornea is not the equivalent in all the three
classes. The outer surface of the cornea images is bright in
colour as compared to that of the normal and post-cataract
images. It is origin of discovering big ring area. Using
Cannys edge detection method edge pixel count (EPC) is
computed and by the computation of EPC, the number of
white pixels in the output of the edge detection is found
out. For normal image, the count is very less and in post
cataract image it is more than the normal where as in the
cataract image it is very large. Object perimeter feature
performed erosion. The normal, cataract and post-cataract
images have a lot of unexpected changes in the gray levels.
SVM classifier is used for classification.
Gao et.al [8], during pre-processing step the images
are transformed into gray channel, so as to facilitate
feature extraction. Texture, homogeneity and intensity
features are extracted. Intensity histogram serves as a key
feature to differentiate transparency from opacity. The
histogram of a lens with cataracts has wider width and the
tail extends to the darker side while, histogram of an
obvious lens normally has slim thickness at clear intensity
level. The other feature extracted is combined texture
information. Here the wavelet coefficients are capable to
describe such texture information. Results from both
anterior and posterior images, representing a combined
wavelet map shall be more appropriate to illustrate
cortical and PSC cataracts. Spatial distribution of intensity
and texture features are also extracted. One confront of
automatic cataract detection is the diversity of the
variance and the opacities of the illuminations in the
images. Information of the total lens image may
demonstrate analogous values for lens images with severe
cataracts and clear lens images. Lens is equally separated
into twenty-four subfields to portray the diverse spatial
variance between them. Then, extract the wavelet
statistics and intensity inside each subfield. SVR classifier
is used to classify the cortical and PSC cataract which is
supervised learning technique, used to train the
regression model.
Xu et.al [9], each lens image is separated into three
sections: nucleus, anterior cortex, and posterior cortex.
Features are extracted from each of the resized sections.
BOF or Bag-of-features extraction is performed. This is
also called as the bag-of words model [10]. BOF model
gives a location-independent global representation of local
character in which properties such as rotation, intensity,
scale or affine invariance can be conserved. Here, the local
features in BOF model are image patches that characterize
texture and intensity information. The local patches from a
set of training images, k-means clustering is used to create
the codebook from arbitrarily selected samples, and the
BOF is obtained in a binning method. A regression model
is used to grade nuclear cataract. A regression model is
trained for the nuclear cataract grading task with the
image feature representation. A condensed representation
could potentially be used, but it is uncertain which colour
channels are most instructive for each section of the lens,
and how many bins is finest for a given channel. A group
sparsity constraint in the regression is applied to choose a
successful subset of the extracted features for nuclear
cataract grading.
Li et.al [11], two computer-aided diagnosis systems for
cataract grading is used. It is based on the landmark
detection using ASM method. Features were extracted

using formerly available clinical work [12, 13]. Six-
dimensional feature was selected and they are: mean
intensity of sulcus, mean intensity inside lens, intensity
ratio between anterior lentils to posterior lentil and colour
on posterior reflex. The last two features were obtained by
visual axis profile analysis and it is the intensity
distribution on a horizontal line through central posterior
reflex. Spoke-like features were used to differentiate
cortical opacities from the posterior sub-capsular
opacities to identify the cortical opacities in ROI in
automatic grading system for cortical. An original image is
renewed to polar coordinate first. Edge detection and local
thresholding were applied in both angular and radial
directions. Region mounting was then applied to detect
the cortical opacities. Angular opacities were subtracted
from radial opacities to maintain only the cortical
opacities as cortical seeds. To eliminate noises as a post-
processing step, size and spatial filters were used. SVM or
Support vector machines regression was employed to
train a grading model and calculate the grade for a testing
image.
Table -1: Comparison of cataract detection techniques
Authors Methods Success
rate (%)
Rafat. R et.al
[5]
Dynamic light scattering
and corneal topography
Not
reported
Nayak et.al[6] Cannys edge detection
method
94
Gao et.al [8] Five-fold cross
validation
51-62
Xu et.al [9] Group sparsity-based
constraint
69
Li et.al [11] Model- based approach,
Thresholding
89.3
3.CONCLUSIONS
Cataract is a familiar crisis in an aging people.
Decreased vision due to cataract can very much influence
the patient's capability to carry out usual actions. Most of
the automatic cataract detection systems are based on
retro illumination images or slit lamp direct images. But
exceptions cannot be handled by using these techniques
while this can be used for mass screening. This work
mainly focused on cataract and its screening methods.
REFERENCES
[1] http://www.aao.org/eye health/diseases/cataracts-
risk
[2] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2698
026/
[3] https://nei.nih.gov/health/cataract/cataract_facts
[4] http://www.allaboutvision.com/smoking/
[5] Rafat R. Ansari National Center for Microgravity
Research, Cleveland, Ohio, Manuel B. Datiles, National
Eye Institute/NIH, Bethesda, Maryland James F. King
Dynacs Engineering Company, Inc., Brook Park, Ohio,
“A New Clinical Instrument for the Early Detection of
Cataract Using Dynamic Light Scattering and Corneal
Topography”.
[6] Jagadish Nayak, “Automated Classification of Normal,
Cataract and Post Cataract Optical Eye Images using
SVM Classifier”, Proceedings of the World Congress on
Engineering and Computer Science 2013 Vol I WCECS
2013, 23-25 October, 2013, San Francisco, USA.
[7] Gonzalez RC, Wintz P, “Digital Image Processing”, 2nd
edn, Addison-Wesley, Reading, MA, 1987.
[8] Xinting Gao, Damon Wing Kee Wong, Tian-Tsong Ng1,
Carol Yim Lui Cheung, Ching-Yu Cheng, and Tien Yin
Wong, “Automatic Grading of Cortical and PSC
Cataracts Using Retroillumination Lens Images”,
Institute for Infocomm Research, A STAR, Singapore,
2009.
[9] Yanwu Xu, Xinting Gao1, Stephen Lin, DamonWing Kee
Wong, Jiang Liu, Dong Xu, Ching-Yu Cheng, Carol Y.
Cheung, and Tien Yin Wong, “Automatic Grading of
Nuclear Cataracts from Slit-Lamp Lens Images Using
Group Sparsity Regression”, in Institute for Infocomm
Research, Agency for Science, Technology and
Research, Singapore, Microsoft Research Asia, P.R.
China, School of Computer Engineering, Nanyang
Technological University, Singapore, Singapore Eye
Research Institute, Singapore.
[10] Fei-Fei, L, Perona, P, “A Bayesian Hierarchical Model
for Learning Natural Scene Categories”. In: CVPR, vol. 2,
pp. 524531 (2005).
[11] Huiqi Li, Joo Hwee Lim, Jiang Liu, Damon Wing Kee
Wong Ngan Meng Tan, Shijian Lu, Zhuo Zhang, Tien Yin
Wong, “Computerized Systems for Cataract Grading”,
Institute for Infocomm Research, A STAR (Agency for
Science, Technology and Research), Singapore and
Singapore Eye Research Institute, Singapore.
[12] L. T. Chylack, J. K. Wolfe, D. M. Singer, M. C. Leske, et al,
“The lens opacities classification system III”, Archives
of Ophthalmology, Vol.111, 1993, pp. 831-836.
[13] B. E. K. Klein, R. Klein, K. L. P. Linton, Y. L. Magli, M. W.
Neider, “Assessment of Cataracts from Photographs in
the Beaver Dam Eye Study,” Ophthalmology, Vo. 97, No.
11, 1990, pp.1428-1433.
[14] Shaohua Fan, Charles R. Dyer, Larry Hubbard, Barbara
Klein “An Automatic System for Classification of
Nuclear Sclerosis from Slit-Lamp Photographs”,
1Department of Computer Science, University of
Wisconsin-Madison, USA, 2003.
[15] Yew Chung Chow, Xinting Gao, Huiqi Li, Joo Hwee Lim,
Ying Sun ,Tien Yin Wong “Automatic Detection of
Cortical and PSC Cataracts Using Texture and Intensity
Analysis on Retro-illumination Lens Images”.

ifferent Techniques for Cataract Detection

More Related Content

What's hot (19)

Similar to ifferent Techniques for Cataract Detection (20)

More from IRJET Journal (20)

Recently uploaded (20)

ifferent Techniques for Cataract Detection