project final ppt
TRANSCRIPT
1
IDENTIFICATION OF DIABETIC RETINOPATHY USING THE RETINAL
IMAGES
Department of ECE
-by M.Ayisha sithika M.Basima banu K.Gayathri
Batch number :111Under the supervision of
Miss L.C.MEENA(Asst.PROFESSOR) ,DACE
2
AIM
To identify diabetic retinopathy using the retinal images in an efficient manner.
Exudates is one of the features used to identify the diabetic retinopathy .
OBJECTIVE
Exudates ,a very important and mostly occurring feature of retinopathy is identified using k-means and naives bayes classifier.
INTRODUCTION-EYE
INTRODUCTION-DIABETIC RETINOPAHTY
DIABETIC RETINOPATHY• DR is an eye disease which has been caused due to high blood sugar
level.
TYPES OF DR
1) Non-proliferative diabetic retinopathy2) Diabetic maculopathy3) Proliferative diabetic retinopathy
Simulation of defective vision as experienced by a Diabetic whose vision has been affected by Diabetic retinopathy
Normal Defective
6
EXUDATES
• Primary sign of diabetic retinopathy • It is lipids and proteins leaks from damaged blood vessels.
FUNDUS IMAGE OF EYE • FUNDUS OF EYE: The back portion of the interior of the eyeball, visible through the pupil by use of the
ophthalmoscope.
• FUNDUS IMAGE: Fundus photography is performed by a fundus camera, which basically consists of a specialized low power microscope with an attached camera.
4
7
LITREATURE SURVEY
1.Akara Sopharak , Bunyarit Uyyanonvara and Sarah Barman[6], “Automatic Exudate Detection from Non-dilated Diabetic Retinopathy-Retinal Images Using Fuzzy C-means Clustering”.
ADVANTAGES: The low contrast retinal image- intensity increased and a number of
edge pixels were extracted.
DISADVANTAGES: More time consuming.
�
8
2.T. Walter, J. Klein, P. Massin, and A. Erginary[2],“A contribution of image processing to the diagnosis of diabetic retinopathy thy,detection of exudates in colour fundus images of the human retina".
ADVANTAGES: Time consumption is reduced as it uses mathematical
morphology techniques .
DISADVANTAGES: The paper ignored some types of errors on the border of the segmented exudates in their reported performances. Time consumption is reduced but not to great extent.
9
NOVELTY USED • In our project we are using k-means clustering algorithm with
naive bayes classifier. • Fuzzy c-means algorithm, as consumes time, so k-means is
used to reduce time. • Naive bayes,a type of classifier is used to increase the
accuracy and sensitivity of the detection.
WORK ACCOMPLISHED
BLOCK DIAGRAM:
INPUT RETINAL IMAGE
PRE-PROCESSING SEGMENTATION FEATURE
EXTRACTION
CLASSIFICATION
EXUDATES
NON-EXUDATES
STEP 1:PRE-PROCESSING
RGB to HIS image
Median filtering
CLAHE
HSI to RGB image
STEP 2:IMAGE SEGMENTATION
RGB to l*a*b colour
space
a*b alone using k-means
clusteringfive clusters
labels every pixel
Colour segmented
images
optic disc is localized
STEP 3:FEATURE EXTRACTION
• On the basis of colour and texture orientation, features are extracted using GLCM.
STEP 4: CLASSIFICATION
• The final step is classification of given input as exudates (or) non-exudates by naive bayes classifier.
PRE-PROCESSED OUTPUT
Input retinal image
H component S Component I Component
HSI Components
Filtered I component CLAHE image Pre-processed image
SEGMENTATION OUTPUT
Image Labeled By Cluster Index
a. L channel b. A channel c. B channel
LAB colour space images
CLUSTER FORMATION
cluster1 cluster2 cluster3
cluster4 cluster5
Cluster Output
EXECUTION OF FINAL OUTPUT
CONCLUSION
• The selected features clustered by k-means clustering and classified into exudates and non –exudates using naive bayes classifier.
• Using this approach, the exudates are detected with 98% success rate.
FUTURE WORK
• Detection of Micro-aneurysm and also maculopathy be predicted and performance can be compared.
REFERENCES[1] Wynne Hsu, P M D S Pallawala, Mong Li Lee, KahGuan Au Eong(2001),”The Role of Domain
Knowledge in the Detection of Retinal Hard Exudates”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Kauai Marriott, Hawaii, vol.12,pp. 533-548.
[2] T. Walter, J.Klein, P.Massin and A.Erginary(2002), “A Contribution of image processing to the
diagnosis of Diabetic Retinopathy detection of exudates in color fundus images of the human retina”, IEEE Trans. On Med. images, vol. 21, no. 10, pp. 1236-1243.
[3] Pizer. S.M(2003),“The Medical Image Display and analysis group at the university of North Carolina:Reminiscences and philosophy ”, IEEE Trans On Medical Imaging, vol. 22, no. 1, pp. 2-10.
[4] Fleming. AD, Philips. S, Goatman. KA, Williams. GJ, Olson.JA, sharp. PF(2007),“Automated detection of exudates for Diabetic Retinopathy Screening”, Journal of Phys. Med. Bio., vol. 52, no. 24, pp. 7385-7396.
[5]Alireza Osareh, Bita Shadgar, and Richard Markham(2009), “A Computational-Intelligence-Based Approach for Detection of Exudates in Diabetic Retinopathy Images”,IEEE Transactions on Information Technology in Biomedicine,vol. 13, no. 4,pp.535-545.International Diabetic Federation (IDF), 2009a, Latest diabetes figures paint grim global picture.
• [6] Akara Sopharak, Bunyarit Uyyanonvara, Sarah Barman(2009), “Automatic Exudate Detection from Non-dilated Diabetic Retinopathy retinal images using Fuzzy Cleans Clustering” Journal of Sensors, vol.9, No. 3, pp 2148- 2161.
• [7] Saiprasad Ravishankar, Arpit Jain, Anurag Mittal(2009),“Automated feature extraction for early detection of Diabetic Retinopathy in fundus images”,IEEE Conference on Computer vision and pattern Recognition, pp. 210-217.
• [8] Doaa Youssef, Nahed Solouma, Amr El-dib, Mai Mabrouk(2010),“New Feature-Based Detection of Blood Vessels and Exudates in Color Fundus Images” IEEE conference on Image Processing Theory, Tools and Applications, vol.16,pp.294-299
• [9] Guoliang Fang, Nan Yang, Huchuan Lu and Kaisong Li(2010),“Automatic Segmentation of Hard Exudates in fundus images based on Boosted Soft Segmentation”, International Conference on Intelligent Control and Information Processing, vol.13,pp. 633-638.
• [10] Plissiti.M.E., Nikar.C, Charchanti.A(2011),“Automated detection of cell nuclei in pap smear images using morphological reconstruction and clustering” IEEE Trans. On Insformation Technology in Biomedicine, vol. 2, pp. 233-241.
Any queries