IRJET- Comparison of Preprocessing Methods for Diabetic Retinopathy Detection using Fundus Images
0 likes42 views
1) The document compares different preprocessing methods for detecting diabetic retinopathy from fundus images, including contrast enhancement, histogram equalization, and Contrast Limited Adaptive Histogram Equalization (CLACHE). 2) It evaluates the performance of these preprocessing techniques on 5 fundus images using metrics like peak signal-to-noise ratio (PSNR), mean squared error (MSE), and entropy. 3) The results show that CLACHE achieved the best performance with higher PSNR and lower MSE values, indicating it succeeds in highly enhancing the fundus images for diabetic retinopathy detection compared to the other methods.
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 252
Comparison of Preprocessing Methods for Diabetic Retinopathy
Detection Using Fundus Images
J. AashikathulZuberiya1, Dr. S. Shajun Nisha2, Dr. M. Mohamed Sathik3
1M.Phil Research Scholar, PG & Research Department of Computer Science, Sadakathullah Appa College,
Tirunelveli, Tamilnadu, India
2Assistant Professor & Head, PG & Research Department of Computer Science, Sadakathullah Appa College,
Tirunelveli, Tamilnadu, India
3Principal, Sadakathullah Appa College, Tirunelveli, Tamilnadu, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - Diabetic Retinopathy is eye disorder among
people with diabetics which may lead to blindness. Diabetes is
a chronic disorder caused by insulin deficiency in the body.
Diabetes for a prolonged time damages the blood vessels of
retina and affects the vision of a person and leads to Diabetic
retinopathy. It is classifies into two categories, non
proliferative diabetic retinopathy (NPDR) and proliferative
diabetic retinopathy (PDR). Fundus photography involves
capturing a photograph of the back of the eye. The rawretinal
fundus images are hard to process. To enhance some features
and to remove unwanted features Preprocessing is used.
Preprocessing techniques like image enhancement, histogram
equalization, Contrast Limited Adaptive Histogram
Equalization (CLACHE) are performed. The results are
evaluated by Mean Square Error (MSE), Peak Signal to Noise
Ratio (PSNR) and Entropy Values.
Key Words: Diabetic Retinopathy, fundus image,
Preprocessing, clache.
1. INTRODUCTION
Diabetes is a chronic disorder caused by the insulin
deficiency body or inability of body cell to respondtoinsulin
in the body. Prolonged Diabetesleadsto manycomplications
like heart disorder, Neuropathy, kidney disorder and eye
disease. The world health organization (WHO)reportedthat
135 million people have diabetes worldwide which may
increase to 300 Million by 2025. Diabetic Retinopathy is an
eye disease among people with diabetics which may lead to
vision impairment or even blindness. It causes loss of vision
in 1.8 Million in 2015 people to 37 Million in 2040. Damage
to the tiny blood vessels in retina from the optic disk inside
the eyes results in Diabetic retinopathy. The anomalies like
micro- anourysms, hemorrhages,hardexudates,cotton wool
spots develops at different stages of diabetic retinopathy.
Diabetic Retinopathy (DR) is classified into Non
Proliferative Diabetic Retinopathy (NPDR) and Proliferative
Diabetic Retinopathy (PDR). Depending on the anomalies or
features present in the retina the stages of the DR can be
identified. NPDR stage has mild, moderate and severe stage.
In NPDR the stage can be ranged from mild , moderate and
severe by the presence of the features in various levels with
less growth of new blood vessels. PDR is an advanced stage
in which the fluids sent by the retina fornourishmenttrigger
the growth of new blood vessel that are abnormal and
fragile. They grow along the retina and along the surface of
vitreous gel which fills inside the eye. It might leak blood
into retina which may result in severe vision loss and even
blindness. Initial stage has no vision problem, but with time
and severity of diabetes it may lead to vision loss[5][6].
Fig -1: Anomalies in human eye
DR can be treated with effective treatments but
there should be early detection and continuous monitoring.
Fundus images are used to diagnosis of DR. Fundus
photography is performed by a fundus camera to record
color images of the condition of the interior surface of the
eye, in order to document the presence of disorders . It
consists of a specialized low power microscope with an
attached camera designedto photographtheinteriorsurface
of the eye, including the retina, retinal vasculature, optic
disc, macula, and the fundus. Patients eyes will be dilated
before the procedure[10].
The raw retinal fundus images are hard to process.
To enhance some features and to removeunwantedfeatures
Preprocessing is used .The pre processing is the important
phase in image processing.. The acquired imageisconverted
into gray scale image. Contrast enhancement, Histogram
equalization, CLACHE are used to improve the quality of
images. Performance of these functions are evaluated and](https://image.slidesharecdn.com/irjet-v6i349-190803102010/85/IRJET-Comparison-of-Preprocessing-Methods-for-Diabetic-Retinopathy-Detection-using-Fundus-Images-1-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 253
compared using metrics to find the results.Theoutlineof the
proposed work is shown in Fig 2.
1.1 Motivation Justification
As there are various preprocessing methodsareavailable
and been introduced, thepurposeofusingPreprocessingisto
enhance somefeaturesand to remove unwantedfeatures.To
identify the best method, Comparision of different image
enhancementareperformed.Basedonthemetricestheresult
is justified.
1.2 Outline of the Paper
Fig -2: Outline of the proposed work.
1.3 Organization of the Paper
Paper continuous as follows - Section II consist of the
related work that is literature survey , Section III describes
methodologies and fundus image enhancement using
different techniques, Section IV is about the experimental
results and discussion Section V has conclusion.
2. RELATED WORK
Swathi.C; Anoop.B.K; D.Anto sahaya dhas; S.Perumal sanker
[1] used different types of preprocessing techniques. They
used adaptive histogram equalization, Weiner filter , Median
filter and adaptive mean filter. PSNR and MSE values are
used for measurement of efficiency.
Sumathy, B.; Poornachandra, S [2] extracted the features of
retinal images using the new adaptivehistogram.Theoptical
disc is difficult to analyze because of its brightness. The
author used the method which gives good results.
Chen Hee Ooi; Isa, N.A.M [3] used two different methods for
adaptive contrast enhancement and brightness presevation.
Using this method they divide the histogram on the basis of
median and uses the advancement of bi- histogram
equalization.
3. METHODOLOGY
Fundus imagesareusedtodiagnosisofDR.Itclassifiesthe
image into normal,NPDRandPDR.Forevaluationfiveimages
were considered[11]. Working with color images makes the
task more complex in image processing, so the color images
are converted to Gray scale Images(GSI).Thencontrastof GSI
image is enhanced to boost the high intensity pixel along
retinal boundaries. The preprocessed output imageisshown
in Table 6.
3.1 Contrast Enhancement
Contrast is an important factor in any subjective
evaluation of image quality. Contrast is created by the
difference in luminance reflectedfromtwoadjacentsurfaces.
In other words, contrast is the difference in visual properties
that makes an object distinguishable from other objects and
the background. In visual perception, contrast is determined
by the difference in the colour and brightness of the object
with other objects.[4]
Low contrast image values concentrated near a
narrow range (mostly dark, or mostly bright, or mostly
medium values). Contrast enhancement (CE) change the
image value change the image value distribution to cover a
wide range. Contrast of an image can be revealed by its
histogram. imadjust function increases the contrast of the
image by mapping the values of the input intensity image to
new values such that, by default, 1% of the data is saturated
at low and high intensities of the input data.
3.2 Histogram Equalization
Histogram equalization is used to enhance contrast.
It is not necessary that contrast will always be increase in
this. There may be some cases were histogram equalization
(HE) can be worse. In that cases the contrast is decreased.
Histeq performs histogram equalization. It enhances the
contrast of images by transforming the values in an intensity
image so that the histogram of the output image
approximately matches a specified histogram (uniform
distribution by default.
3.3 Contrast Limited Adaptive Histogram
Equalization
Contrast-Limited Adaptive Histogram (CLACHE) uses
adapthisteq function to performs equalization. It enhances
the contrast of the grayscale image by transforming the
values using contrast-limited adaptive histogram
equalization. Unlike histogram equalization, it operates on
small data regions (tiles) rather than the entire image. Each
tile's contrast is enhanced so that the histogram of each
outputregionapproximatelymatchesthespecifiedhistogram
(uniform distribution by default). The contrastenhancement
can be limited in order to avoid amplifying the noise which
might be present in the image[7]](https://image.slidesharecdn.com/irjet-v6i349-190803102010/85/IRJET-Comparison-of-Preprocessing-Methods-for-Diabetic-Retinopathy-Detection-using-Fundus-Images-2-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 254
4. EXPERIMENTAL RESULTS
4.1 Performance Metrics
The performance metrics such as Peak signal to Noise Ratio
(PSNR), Mean Squared Error (MSE), Entropy are
calculated.
A. Peak Signal To Noise Ratio (PSNR):
It is the evaluation standard of the reconstructed image
quality, is the most wanted feature. PSNR is measured in the
decibels (dB) and it is given by
PSNR = 10log (255 ∕ 2 MSE)
Where the value 255 is the maximum possible value that can
be attained by the image signal. Mean square error is defined
as where M*N is the size of the original image. Higher the
PSNR value betters the reconstructed image.
B. Mean Square Error (MSE):
The average squared difference betweenthereferencesignal
and distorted signal is called as the mean square error. It can
be easily calculated by adding up the squared difference
pixel-by-pixel and dividing by the total pixel count. Let m x n
is a noise free monochrome image I, and K is defined as the
noisy approximation. Then the mean square error between
these two signals is defined as
MSE = 1 m × n [ I i, j − K i, j ] n − 1 2 j = 0 m − 1 i = 0
C. Entropy:
For a given PDF p, entropy Ent[p] is computed. In
general, the entropy is a useful tool to measure the richness
of the details in the output image.
Ent[p] = −Σk = 0 (k)log2 p (k)
Table -1: Image Quality Parameters of Image 1
Image 1 MSE PSNR Entropy
CE 6546.24 10.00 5.8396
HE 3662.97 12.53 5.4481
CLACHE 411.15 22.02 6.4827
Table -2: Image Quality Parameters of Image 2
Image 2 MSE PSNR Entropy
CE 4780.83 11.37 6.1269
HE 2470.17 14.24 5.6041
CLACHE 437.50 21.75 6.7184
Table -3: Image Quality Parameters of Image 3
Image 3 MSE PSNR Entropy
CE 9615.47 8.34 5.6421
HE 4306.52 11.82 5.2647
CLACHE 366.68 22.52 6.1634
Table -4: Image Quality Parameters of Image 4
Image 4 MSE PSNR Entropy
CE 3834.98 12.33 6.2004
HE 4492.04 11.64 5.6039
CLACHE 680.44 19.84 6.7073
Table -5: Image Quality Parameters of Image 5
Image 5 MSE PSNR Entropy
CE 7613.48 9.35 5.9684
HE 4859.93 11.30 5.5001
CLACHE 708.94 19.66 6.5956
4.2 Performance Evaluation
In this proposed work preprocessing techniques for retinal
fundus images were applied for fiveimages.Techniquessuch
as gray scale image. Contrast enhancement, Histogram
equalization, CLACHE are applied and the quality of images
are improved. The results of the images are tabulated in
Table 6 and Table 7 .
Table -6: Result Of Images on applying Gray scale and
Contrast Enhancement.
ImageName Input Image GSI CE
Image 1
Image 2
Image 3
Image 4
Image 5](https://image.slidesharecdn.com/irjet-v6i349-190803102010/85/IRJET-Comparison-of-Preprocessing-Methods-for-Diabetic-Retinopathy-Detection-using-Fundus-Images-3-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 255
Table -6: Result of images on applying Histogram
Equalization and CLACHE.
5. CONCLUSIONS
In this paper preprocessing techniques for retinal fundus
images were applied for five images. The image quality
obtained after applying these algorithms is assessed with
metrics. These metrics include Peak Signal To Noise Ratio
(PSNR), Entropy and Mean Square Error (MSE). From the
results Contrast Limited Adaptive Histogram Equalization
(CLACHE) succeeds because it has higher PSNR and Lower
MSE value. Using CLACHE will results in highly enhanced
image.
REFERENCES
[1] Swathi C, Anoop B.K, Anto Sahaya Dhas D, Perumal
Sanker S, “Comparison OfDifferentImagePreprocessing
Methods Used For Retinal Fundus Images”, Proc.IEEE
Conference On Emerging Devices And Smart
Systems(ICEDSS 2017) 3-4march2017 Vol.,No.,PP.978-
1-5090-5555-5/17$31.00
[2] Sumathy B, Poornachandra S, “Feature Extraction In
Retinal Fundus Images”, Information Communication
And Embedded Systems (ICICES), 2013 International
Conference On , vol., no., pp.798-802, 21-22 Feb. 2013
[3] Chen Hee Ooi; Isa, N.A.M., “Adaptive Contrast
Enhancement Methods With Brightness Preserving”,
Consumer Electronics, IEEE Transactions On , vol.56,
no.4, pp.2543-2551, November 2010
[4] Preeti Gupta, “Contrast Enhancemet For Retinal Images
Using Multi-Objective Genetic Algorithm”, International
Journal Of Emerging Trends In Engineering And
Development, Issue 6, Vol. 1 (January 2016) ISSN 2249-
6149.
[5] Swati gupta and Karandikar,”Diagonosis of diabetic
retinopathy using machine learning”, Journal of
Research and development, 2015
[6] Sarni Suhaila Rahim, Vasile Palade, Chrisina Jayne,
Andreas Holzinger, James Shuttle worth, “Detection of
Diabetic Retinopathy and Masculopathy in Eye Fundus
Images using Fuzzy Image Processing”, Springer
International Publishing Switzerland 2015, Y.Guo et al.
(Eds.): BIH 2015 , LNAI 9250, pp.379-388, 2015
[7] Jeline Devadhas and R. Binisha, “Early Diagonosis of
Diabetic Retinopathy by the Detection of
Microaneurysms in Fundus Images”, ICTACT Journal On
Image and Video Processing , ISSN: 0976-9102,
Volume:08, Issue: 04
[8] G. Prabavathi, Dr.K.Mahesh, “Automated Analysis of
Microaneurysm Detection of Diabetic Retinopathy”,
International Journal for Modern Trends in Science and
Technology, ISSN: 2455-3778, Volume:03, Issue No: 05,
May 2017
[9] Salman Sayed, Dr. Vandana Inamdar, Sangram Kapre, “
Detection of Diabetic Retinopathy usimg Image
ProcessingandMachineLearning”,International Journal
of Innovative Research in Science, Engineering and
Technology, ISSN: 2319-8753, Vol.6, Issue 1, January
2017
[10] Khan Abdul Mukhtar Khan, “Detection of Digital Image
Processing A Survey”, International Journal of Research
in Computer Application and Robotics”,ISSN 2320-7345
Vol.6 Issue 1, Pg:13-19, January 2018
[11] Shajun Nisha, Ashika Parvin, “Preservation Of
Historical Document Using Enhancement Techniques
International Journal of Trend in Research and
Development, Vol 4(1), ISSN: 2394-9333 Jan- Feb
2016
BIOGRAPHIES
J.AashikathulZuberiya, M.Phil Research
scholar, currently pursuing at
sadakathullah appa college, Tirunelveli. I
had completed my UG & PG degree in
Computer science at Sadakathullah Appa
College, I had certification of NPTEL
courses. My research area is in Image
processing.
Dr.S.ShajunNisha,AssistantProfessorand
Head of the PG & Research Department of
ComputerScience, Sadakathullah Appa
College, Tirunelveli. She has completed
M.Phil. (Computer Science) M.Tech
(Computer and Information Technology)
in Manonmaniam Sundaranar University,
Tirunelveli and completed Ph.D (Computer Science) in
Bharathiyar university,Coimbatore. She has involved in
various academic activities and attended so many national
and international seminars, conferences and presented
ImageName HE CLACHE
Image 1
Image 2
Image 3
Image 4
Image 5](https://image.slidesharecdn.com/irjet-v6i349-190803102010/85/IRJET-Comparison-of-Preprocessing-Methods-for-Diabetic-Retinopathy-Detection-using-Fundus-Images-4-320.jpg)
IRJET- Comparison of Preprocessing Methods for Diabetic Retinopathy Detection using Fundus Images
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 252 Comparison of Preprocessing Methods for Diabetic Retinopathy Detection Using Fundus Images J. AashikathulZuberiya1, Dr. S. Shajun Nisha2, Dr. M. Mohamed Sathik3 1M.Phil Research Scholar, PG & Research Department of Computer Science, Sadakathullah Appa College, Tirunelveli, Tamilnadu, India 2Assistant Professor & Head, PG & Research Department of Computer Science, Sadakathullah Appa College, Tirunelveli, Tamilnadu, India 3Principal, Sadakathullah Appa College, Tirunelveli, Tamilnadu, India ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract - Diabetic Retinopathy is eye disorder among people with diabetics which may lead to blindness. Diabetes is a chronic disorder caused by insulin deficiency in the body. Diabetes for a prolonged time damages the blood vessels of retina and affects the vision of a person and leads to Diabetic retinopathy. It is classifies into two categories, non proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). Fundus photography involves capturing a photograph of the back of the eye. The rawretinal fundus images are hard to process. To enhance some features and to remove unwanted features Preprocessing is used. Preprocessing techniques like image enhancement, histogram equalization, Contrast Limited Adaptive Histogram Equalization (CLACHE) are performed. The results are evaluated by Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Entropy Values. Key Words: Diabetic Retinopathy, fundus image, Preprocessing, clache. 1. INTRODUCTION Diabetes is a chronic disorder caused by the insulin deficiency body or inability of body cell to respondtoinsulin in the body. Prolonged Diabetesleadsto manycomplications like heart disorder, Neuropathy, kidney disorder and eye disease. The world health organization (WHO)reportedthat 135 million people have diabetes worldwide which may increase to 300 Million by 2025. Diabetic Retinopathy is an eye disease among people with diabetics which may lead to vision impairment or even blindness. It causes loss of vision in 1.8 Million in 2015 people to 37 Million in 2040. Damage to the tiny blood vessels in retina from the optic disk inside the eyes results in Diabetic retinopathy. The anomalies like micro- anourysms, hemorrhages,hardexudates,cotton wool spots develops at different stages of diabetic retinopathy. Diabetic Retinopathy (DR) is classified into Non Proliferative Diabetic Retinopathy (NPDR) and Proliferative Diabetic Retinopathy (PDR). Depending on the anomalies or features present in the retina the stages of the DR can be identified. NPDR stage has mild, moderate and severe stage. In NPDR the stage can be ranged from mild , moderate and severe by the presence of the features in various levels with less growth of new blood vessels. PDR is an advanced stage in which the fluids sent by the retina fornourishmenttrigger the growth of new blood vessel that are abnormal and fragile. They grow along the retina and along the surface of vitreous gel which fills inside the eye. It might leak blood into retina which may result in severe vision loss and even blindness. Initial stage has no vision problem, but with time and severity of diabetes it may lead to vision loss[5][6]. Fig -1: Anomalies in human eye DR can be treated with effective treatments but there should be early detection and continuous monitoring. Fundus images are used to diagnosis of DR. Fundus photography is performed by a fundus camera to record color images of the condition of the interior surface of the eye, in order to document the presence of disorders . It consists of a specialized low power microscope with an attached camera designedto photographtheinteriorsurface of the eye, including the retina, retinal vasculature, optic disc, macula, and the fundus. Patients eyes will be dilated before the procedure[10]. The raw retinal fundus images are hard to process. To enhance some features and to removeunwantedfeatures Preprocessing is used .The pre processing is the important phase in image processing.. The acquired imageisconverted into gray scale image. Contrast enhancement, Histogram equalization, CLACHE are used to improve the quality of images. Performance of these functions are evaluated and
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 253 compared using metrics to find the results.Theoutlineof the proposed work is shown in Fig 2. 1.1 Motivation Justification As there are various preprocessing methodsareavailable and been introduced, thepurposeofusingPreprocessingisto enhance somefeaturesand to remove unwantedfeatures.To identify the best method, Comparision of different image enhancementareperformed.Basedonthemetricestheresult is justified. 1.2 Outline of the Paper Fig -2: Outline of the proposed work. 1.3 Organization of the Paper Paper continuous as follows - Section II consist of the related work that is literature survey , Section III describes methodologies and fundus image enhancement using different techniques, Section IV is about the experimental results and discussion Section V has conclusion. 2. RELATED WORK Swathi.C; Anoop.B.K; D.Anto sahaya dhas; S.Perumal sanker [1] used different types of preprocessing techniques. They used adaptive histogram equalization, Weiner filter , Median filter and adaptive mean filter. PSNR and MSE values are used for measurement of efficiency. Sumathy, B.; Poornachandra, S [2] extracted the features of retinal images using the new adaptivehistogram.Theoptical disc is difficult to analyze because of its brightness. The author used the method which gives good results. Chen Hee Ooi; Isa, N.A.M [3] used two different methods for adaptive contrast enhancement and brightness presevation. Using this method they divide the histogram on the basis of median and uses the advancement of bi- histogram equalization. 3. METHODOLOGY Fundus imagesareusedtodiagnosisofDR.Itclassifiesthe image into normal,NPDRandPDR.Forevaluationfiveimages were considered[11]. Working with color images makes the task more complex in image processing, so the color images are converted to Gray scale Images(GSI).Thencontrastof GSI image is enhanced to boost the high intensity pixel along retinal boundaries. The preprocessed output imageisshown in Table 6. 3.1 Contrast Enhancement Contrast is an important factor in any subjective evaluation of image quality. Contrast is created by the difference in luminance reflectedfromtwoadjacentsurfaces. In other words, contrast is the difference in visual properties that makes an object distinguishable from other objects and the background. In visual perception, contrast is determined by the difference in the colour and brightness of the object with other objects.[4] Low contrast image values concentrated near a narrow range (mostly dark, or mostly bright, or mostly medium values). Contrast enhancement (CE) change the image value change the image value distribution to cover a wide range. Contrast of an image can be revealed by its histogram. imadjust function increases the contrast of the image by mapping the values of the input intensity image to new values such that, by default, 1% of the data is saturated at low and high intensities of the input data. 3.2 Histogram Equalization Histogram equalization is used to enhance contrast. It is not necessary that contrast will always be increase in this. There may be some cases were histogram equalization (HE) can be worse. In that cases the contrast is decreased. Histeq performs histogram equalization. It enhances the contrast of images by transforming the values in an intensity image so that the histogram of the output image approximately matches a specified histogram (uniform distribution by default. 3.3 Contrast Limited Adaptive Histogram Equalization Contrast-Limited Adaptive Histogram (CLACHE) uses adapthisteq function to performs equalization. It enhances the contrast of the grayscale image by transforming the values using contrast-limited adaptive histogram equalization. Unlike histogram equalization, it operates on small data regions (tiles) rather than the entire image. Each tile's contrast is enhanced so that the histogram of each outputregionapproximatelymatchesthespecifiedhistogram (uniform distribution by default). The contrastenhancement can be limited in order to avoid amplifying the noise which might be present in the image[7]
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 254 4. EXPERIMENTAL RESULTS 4.1 Performance Metrics The performance metrics such as Peak signal to Noise Ratio (PSNR), Mean Squared Error (MSE), Entropy are calculated. A. Peak Signal To Noise Ratio (PSNR): It is the evaluation standard of the reconstructed image quality, is the most wanted feature. PSNR is measured in the decibels (dB) and it is given by PSNR = 10log (255 ∕ 2 MSE) Where the value 255 is the maximum possible value that can be attained by the image signal. Mean square error is defined as where M*N is the size of the original image. Higher the PSNR value betters the reconstructed image. B. Mean Square Error (MSE): The average squared difference betweenthereferencesignal and distorted signal is called as the mean square error. It can be easily calculated by adding up the squared difference pixel-by-pixel and dividing by the total pixel count. Let m x n is a noise free monochrome image I, and K is defined as the noisy approximation. Then the mean square error between these two signals is defined as MSE = 1 m × n [ I i, j − K i, j ] n − 1 2 j = 0 m − 1 i = 0 C. Entropy: For a given PDF p, entropy Ent[p] is computed. In general, the entropy is a useful tool to measure the richness of the details in the output image. Ent[p] = −Σk = 0 (k)log2 p (k) Table -1: Image Quality Parameters of Image 1 Image 1 MSE PSNR Entropy CE 6546.24 10.00 5.8396 HE 3662.97 12.53 5.4481 CLACHE 411.15 22.02 6.4827 Table -2: Image Quality Parameters of Image 2 Image 2 MSE PSNR Entropy CE 4780.83 11.37 6.1269 HE 2470.17 14.24 5.6041 CLACHE 437.50 21.75 6.7184 Table -3: Image Quality Parameters of Image 3 Image 3 MSE PSNR Entropy CE 9615.47 8.34 5.6421 HE 4306.52 11.82 5.2647 CLACHE 366.68 22.52 6.1634 Table -4: Image Quality Parameters of Image 4 Image 4 MSE PSNR Entropy CE 3834.98 12.33 6.2004 HE 4492.04 11.64 5.6039 CLACHE 680.44 19.84 6.7073 Table -5: Image Quality Parameters of Image 5 Image 5 MSE PSNR Entropy CE 7613.48 9.35 5.9684 HE 4859.93 11.30 5.5001 CLACHE 708.94 19.66 6.5956 4.2 Performance Evaluation In this proposed work preprocessing techniques for retinal fundus images were applied for fiveimages.Techniquessuch as gray scale image. Contrast enhancement, Histogram equalization, CLACHE are applied and the quality of images are improved. The results of the images are tabulated in Table 6 and Table 7 . Table -6: Result Of Images on applying Gray scale and Contrast Enhancement. ImageName Input Image GSI CE Image 1 Image 2 Image 3 Image 4 Image 5
- 4. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 255 Table -6: Result of images on applying Histogram Equalization and CLACHE. 5. CONCLUSIONS In this paper preprocessing techniques for retinal fundus images were applied for five images. The image quality obtained after applying these algorithms is assessed with metrics. These metrics include Peak Signal To Noise Ratio (PSNR), Entropy and Mean Square Error (MSE). From the results Contrast Limited Adaptive Histogram Equalization (CLACHE) succeeds because it has higher PSNR and Lower MSE value. Using CLACHE will results in highly enhanced image. REFERENCES [1] Swathi C, Anoop B.K, Anto Sahaya Dhas D, Perumal Sanker S, “Comparison OfDifferentImagePreprocessing Methods Used For Retinal Fundus Images”, Proc.IEEE Conference On Emerging Devices And Smart Systems(ICEDSS 2017) 3-4march2017 Vol.,No.,PP.978- 1-5090-5555-5/17$31.00 [2] Sumathy B, Poornachandra S, “Feature Extraction In Retinal Fundus Images”, Information Communication And Embedded Systems (ICICES), 2013 International Conference On , vol., no., pp.798-802, 21-22 Feb. 2013 [3] Chen Hee Ooi; Isa, N.A.M., “Adaptive Contrast Enhancement Methods With Brightness Preserving”, Consumer Electronics, IEEE Transactions On , vol.56, no.4, pp.2543-2551, November 2010 [4] Preeti Gupta, “Contrast Enhancemet For Retinal Images Using Multi-Objective Genetic Algorithm”, International Journal Of Emerging Trends In Engineering And Development, Issue 6, Vol. 1 (January 2016) ISSN 2249- 6149. [5] Swati gupta and Karandikar,”Diagonosis of diabetic retinopathy using machine learning”, Journal of Research and development, 2015 [6] Sarni Suhaila Rahim, Vasile Palade, Chrisina Jayne, Andreas Holzinger, James Shuttle worth, “Detection of Diabetic Retinopathy and Masculopathy in Eye Fundus Images using Fuzzy Image Processing”, Springer International Publishing Switzerland 2015, Y.Guo et al. (Eds.): BIH 2015 , LNAI 9250, pp.379-388, 2015 [7] Jeline Devadhas and R. Binisha, “Early Diagonosis of Diabetic Retinopathy by the Detection of Microaneurysms in Fundus Images”, ICTACT Journal On Image and Video Processing , ISSN: 0976-9102, Volume:08, Issue: 04 [8] G. Prabavathi, Dr.K.Mahesh, “Automated Analysis of Microaneurysm Detection of Diabetic Retinopathy”, International Journal for Modern Trends in Science and Technology, ISSN: 2455-3778, Volume:03, Issue No: 05, May 2017 [9] Salman Sayed, Dr. Vandana Inamdar, Sangram Kapre, “ Detection of Diabetic Retinopathy usimg Image ProcessingandMachineLearning”,International Journal of Innovative Research in Science, Engineering and Technology, ISSN: 2319-8753, Vol.6, Issue 1, January 2017 [10] Khan Abdul Mukhtar Khan, “Detection of Digital Image Processing A Survey”, International Journal of Research in Computer Application and Robotics”,ISSN 2320-7345 Vol.6 Issue 1, Pg:13-19, January 2018 [11] Shajun Nisha, Ashika Parvin, “Preservation Of Historical Document Using Enhancement Techniques International Journal of Trend in Research and Development, Vol 4(1), ISSN: 2394-9333 Jan- Feb 2016 BIOGRAPHIES J.AashikathulZuberiya, M.Phil Research scholar, currently pursuing at sadakathullah appa college, Tirunelveli. I had completed my UG & PG degree in Computer science at Sadakathullah Appa College, I had certification of NPTEL courses. My research area is in Image processing. Dr.S.ShajunNisha,AssistantProfessorand Head of the PG & Research Department of ComputerScience, Sadakathullah Appa College, Tirunelveli. She has completed M.Phil. (Computer Science) M.Tech (Computer and Information Technology) in Manonmaniam Sundaranar University, Tirunelveli and completed Ph.D (Computer Science) in Bharathiyar university,Coimbatore. She has involved in various academic activities and attended so many national and international seminars, conferences and presented ImageName HE CLACHE Image 1 Image 2 Image 3 Image 4 Image 5
- 5. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 03 | Mar 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 256 numerous research papers. She is a member of ISTE and IEANG and her specialization is Imageprocessingand neural network. Dr.M.Mohamed Sathik, Principal Sadakathullah Appa College, Tirunelveli. He has completed Ph.D(Computer science & engineering) Ph.D(Computer science), M.Phil. (Computer Science), M.Tech(Computer Science and Information Technology) in Manonmaniam Sundaranar University, Tirunelveli. He has so far guided more than 35 research scholars. He has published more than 100 papers in International Journals and also two books.Heisa memberof curriculum development committee of various universities and autonomous colleges of Tamil Nadu. He is a syndicate member Manonmaniam Sundaranar University, Tirunelveli. His specializations are VRML, Image Processing and Sensor Networks.