Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Thesis Defense

No description
by

Arslan Ahmad

on 3 April 2016

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Thesis Defense

Thank You
OUTLINE
Introduction
Motivation
Literature Review
Proposed Methodology for Blood Vessel Extraction
Proposed Methodology for exudate localization
Results
Conclusion
Future work
Introduction
MOTIVATION
OBJECTIVE
Automated detection and screening of diabetic retinopathy
LITERATURE REVIEW
Automated Detection and Screening of Diabetic Retinopathy (DR) using Fundus Image Analysis
PRESENTED BY:
ARSLAN AHMAD

PRINCIPAL INVESTIGATOR

DR. RAFIA MUMTAZ

WHAT IS DIABETIC RETINOPATHY?
Microvascular complication of diabetes
Thin retinal blood vessel burst
Contamination of retina
Result: Vision impairment
Asymptomatic
RISK FACTOR
Diabetes a global problem
Estimated global prevalence of diabetes
Rise from 2.8 in 2000 to 4.4 million in 2030*








*Wild S, Roglic G, Green A, Sicree R, King H., Global prevalence of diabetes: estimates for the year 2000 and projections for 2030, Diabetes Care,
1047-1053, 27, 2004.

Global Prevalence of Diabetes in million

Probability of diabetic retinopathy**
99% for Type I diabetes
66% for Type II diabetes
Complete vision loss






**Klein, Ronald, Barbara Ek Klein, and SCOT E. Moss, Visual impairment in diabetes,
Ophthalmology 91 (1984): l-9


RISK FACTOR

Probability of DR

Red Lesions
Microaneurysm
Hemorrhages
White Lesions
Exudates

Signs of DR

Non availability of Specialized ophthalmologists
Cost
Asymptomatic

MOTIVATION

Active area of research
Increasing trend in research articles in last three years

Literature Review

DATABASES FOR DR
Private databases
Publicly available databases
DRIVE database
STARE database
ImageRET
DIARETDB0
DIARETDB1
Messidor
Retinopathy Online Challenge

COMPUTATIONAL STEPS IN DR
Detection of DR lies in domain of:
Image processing
Pre-processing
Segmentation
Pattern classification
Feature Extraction
Classification

Green Channel
Image Normalization
Histogram Equalization
Correction of
Non Uniform Illumination
Morphological Operations

PREPROCESSING

Classifiers used in published articles in past three years:
Support Vector Machine
K-Nearest Neighbor
Gaussian Model Mixer
Naïve Bayes
Bayesian
Neural Network




CLASSIFICATION OF MA’S, HA’S,
EXUDATES & COTTON WOOLEN SPOTS

PRE PROCESSING
OPTIC DISK LOCALIZATION & SEGMENTATION
LOCALIZATION OF EXUDATES
PROPOSED METHODOLOGY FOR EXUDATES LOCALIZATION
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Green Channel Extraction
Maximum contrast to distinguish between retinal features
Image Normalization
To enhance the contrast
To remove non uniform illumination
Suppose
I(x,y) be Input Image then estimated normalized image will be given by following equation:
Image Normalization
Morphological closing
Morphological Closing operation was performed to remove blood vessels
Proposed Methodology for Blood Vessel Extraction
PRE PROCESSING
Green channel extraction
CLASSIFICATION
Neural network Classifier
196 input layer
20 hidden layers
binary output
FEATURE EXTRACTION
Gabor wavelet based feature vector
PERFORMANCE MEASURES
Results are evaluated on the basis of following performance measures:
Receiver Operating Characteristics
Area under curve (AUC) in ROC
Sensitivity
Specificity
PERFORMANCE OF PROPOSED SYSTEM FOR EXUDATE LOCALIZATION AND DETECTION
AUC=0.94
Sensitivity=93.91%
Specificity=75.9%
PERFORMANCE OF BLOOD VESSEL EXTRACTION
AUC=0.95
Sensitivity=94.5%
Specificity=78.3%
Conclusion
The proposed system can be used for automated detection and screening of DR
Has ability
Preprocessing
Blood Vessel Extraction
Optic Disk Segmentation
Localization and Detection of Exudates
Risk Factor
Anatomical Structure of Retina
Frequency of Published Articles in Past Three Years
Frequency Distribution of Databases found in Literature
Number of Articles
Frequency Distribution of Preprocessing Techniques
INPUT IMAGE
23
25
26
28
Block Diagram of Proprocessing Stage
30
31
33
*Wild S, Roglic G, Green A, Sicree R, King H., Global prevalence of diabetes: estimates for the year 2000 and projections for 2030, Diabetes Care,
1047-1053, 27, 2004.

**Klein, Ronald, Barbara Ek Klein, and SCOT E. Moss, Visual impairment in diabetes,
Ophthalmology 91 (1984): l-9
INPUT IMAGE
OUTPUT IMAGE
GABOR WAVELET FILTER
Directional Filter
Applied on Multiple Scales
3 to 7 with increment of 2
Multiple Orientations
0 to 165 Degrees with increment of 15 degree
Multiple Aspect ratio
1 to 2 with increment of 0.5
Filter bank comprise of 108 gabor filter
Gabor filters in different orientations starting from top left 0 to 165 degree

L2 Norm
For superposition of all filters responses

Gabor Wavelet
Questions?
Comparison
Future Work
The detect of diabetic Retinopathy by Microaneurysm and Hemorrhages
[1]H. F. Jelinek, R. Pires, R.Padilha, S. Goldenstein, J. Wainer, T. Bossomaier,
and A. Rocha, “Data fusion for multi-lesion diabetic retinopathy detection,” in
Computer-Based Medical Systems (CBMS), 2012 25th International Symposium
on. IEEE, 2012, pp. 1–4.
[2]A. Rocha, T. Carvalho, H. F. Jelinek, S. Goldenstein, and J. Wainer, “Points of
interest and visual dictionaries for automatic retinal lesion detection,” Biomedical
Engineering, IEEE Transactions on, vol. 59, no. 8, pp. 2244–2253, 2012.
[3] L. Giancardo, F. Meriaudeau, T. P. Karnowski, Y. Li, K. W. Tobin, and E. Chaum,
“Automatic retina exudates segmentation without a manually labelled training
set,” in Biomedical Imaging: From Nano to Macro, 2011 IEEE International
Symposium on. IEEE, 2011, pp. 1396–1400.
Full transcript