Scribd
Upload
765 views23 pages

Automated Diagnosis of Diabetic Retinopathy Using Neural Networks

The document discusses a proposed system to automate the diagnosis of diabetic retinopathy using neural networks. It aims to classify retinopathy into non-proliferative and proliferative stages and determine the appropriate treatment. The system architecture includes modules for data construction, preparation, diagnosis using neural networks, and model deployment in clinical practice. The proposed solution aims to reduce manual effort and allow for early prediction and treatment of retinopathy.

Uploaded by

samreen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
765 views23 pages

Automated Diagnosis of Diabetic Retinopathy Using Neural Networks

The document discusses a proposed system to automate the diagnosis of diabetic retinopathy using neural networks. It aims to classify retinopathy into non-proliferative and proliferative stages and determine the appropriate treatment. The system architecture includes modules for data construction, preparation, diagnosis using neural networks, and model deployment in clinical practice. The proposed solution aims to reduce manual effort and allow for early prediction and treatment of retinopathy.

Uploaded by

samreen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 23

Automated Diagnosis of Diabetic Retinopathy

Using Neural Networks.

PRESENTED BY:
ATIKA AFREEN
1604-19-742-016
CONTENTS

 ABSTRACT  WORKING
 OBJECTIVE  RESULTS
 INTRODUCTION  CONCLUSION
 EXISTING SYSTEM  FUTURE ENHANCEMENT
 DRAWBACKS  REFERENCES
 PROPOSED SYSTEM
 ADVANTAGES
 LITERATURE SURVEY
 SYSTEM ARCHITECTURE
ABSTRACT

 Retinopathy is any damage to the retina of eyes which results in


vision impairment and usually occurs in diabetic patients.

 Early detection and diagnosis for retinopathy is essential.


 Hence in order to automate this process with proper treatment methods,
system presents a dataset of DR fundus images and provides a detection
system using neural networks and cloud computing platforms.

OBJECTIVE

 The objective of this system is to automate the diagnosis of DR .


INTRODUCTION

 To improve the retinopathy diagnosis procedures in medical field


yields with fewer medical resources across different areas and
locations.

 Substantial innovations have been made through out medical field.

 To respond these lack of facilities with a range of issues, retinopathy


diagnosis needs the proper treatment and detection methods.
 Nueral networks and cloud computing are one such platforms that can
automate the diagnosis of retinopathy in various medical fields .

 It’s accurate results brings out a wider invention in diagnosis


procedures.
EXISTING SYSTEM

 There were several works carried out on diagnosis of retinopathy


using image processing and other techniques which were not much
effective and didn't bought the automation, accuracy in results.
DRAWBACKS:

1. There was no automatic detection system.

2. Lots of manual efforts.

3. Didn't benefit the diabetic patients in aspects of early diagnosis and


detection of retinopathy,
PROPOSED SYSTEM

 System provides the users with the early diagnosis which predicts the
stage and level associated with retinopathy based on fundus images and
determines the treatment required for it.
ADVANTAGES

 Reduces the manual efforts.

 Early prediction.

 Can be used by the people in the areas where there are no specialists
available.

 Easy to use.
LITERATURE SURVEY

 Rubini and kunthavai carried there work in classifying stages of


retinopathy using machine learning algorithms.

 Mookiah proposed a system that used hybrid features and entropy for
classification of retinopathy.

 Bhatkar and kharat explored the use of multilayer perception neural


network to enhance the diagnosis by classifying retinopathy.
PROPOSED SOLUTION

 The system aims to automate the diagnosis of retinopathy using neural


networks .

 Retinopathy is classified into following two stages :

 Non proliferative DR

 Proliferative DR (It can be classified as mild PDR, moderate PDR, severe


PDR)
SYSTEM ARCHITECTURE

System architecture is classified into four modules as follows:


1. Data Construction
2. Data Preparation
3. DR Diagnosis with Neural Network
4. Model Deployment And Clinical Practise
DATASET CONSTRUCTION
DATA PREPARATION
DR DIAGNOSIS WITH NEURAL NETWORKS
MODEL DEPLOYMENT AND CLINICAL PRACTISE
START

WORKING
LOGIN TO THE WEB SERVICE

UPLOAD THE FUNDUS IMAGES

SYSTEM DETECTS THE STAGE


AND LEVEL OF RETINOPATHY
WITH DATA AVAILABLE ON
CLOUD

PDR / NO
NPDR GENERATES REPORT AS NORMAL
?
YES
STOP
GENERATES REPORT WITH THE
ASSOCIATED LEVEL OF NPDR/ PDR

STOP
RESULT ANALYSIS
FUTURE ENHANCEMENT

 Electronic Health Records of diabetic patients can be used in order


to achieve more accuracy and faster results.

 EHR plays a essential role in reducing the tasks and making it


simpler and easier to achieve.
CONCLUSION

 Automated diabetic retinopathy diagnosis is more advantageous in


several aspects and helps the patients to detect the stages of it at the
earliest followed by the corresponding treatment.
REFERENCES

[1] A. D. Hoover, V. Kouznetsova, M. Goldbaum, "Locating blood vessels in retinal images by


piecewise threshold probing of a matched filter response", IEEE Trans. Med. Imag., vol. 19, no.
3, pp. 203-210, Mar. 2000.
[2] K. Simonyan, A. Zisserman, Very deep convolutional networks for large-scale image
recognition, Sep. 2014, Available: https://arxiv.org/abs/1409.1556.
[3] M. R. K. Mookiah et al., "Evolutionary algorithm based classifier parameter tuning for
automatic diabetic retinopathy grading: A hybrid feature extraction approach", Knowl.-Based
Syst.,vol.39,pp.9-22,Feb.2013,
Available: http://www.sciencedirect.com/science/article/pii/S0950705112002651.
[4] M. D. Bloice, C. Stocker, A. Holzinger, Augmentor: An image augmentation library for
machine learning, Aug. 2017, Available: https://arxiv.org/abs/1708.04680.
[5] S. Resnikoff, W. Felch, T.-M. Gauthier, B. Spivey, "The number of ophthalmologists in
practice and training worldwide: A growing gap despite more than 200 000 practitioners", Brit. J.
Ophthalmol., vol. 96, no. 6, pp. 783-787, 2012, Available: http://bjo.bmj.com/content/96/6/783.

You might also like