Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 13

Eye Disease Prediction Using CNN

Team Members:
Sweta Maharjan (24175/076)
Kritagya Man Shakya (24147/076)
Hardik Dahal (24143/076)
Introduction

 Eye Disease Prediction Application is an application used to


predict a eye disease like Cataract and Glaucoma based on the
ocular image of user.
 Early detection and timely intervention are crucial for
preventing irreversible damage and preserving vision.
Problem Statement

 Eye diseases contributes to vision impairment and blindness if


left undetected and untreated.
 Traditional methods of eye disease detection rely heavily on
clinical examination by trained ophthalmologists which can be
resource-intensive and time consuming.
Objectives

 To harness the power of Deep Learning to predict eye diseases.


 To contribute to the ongoing advancements in healthcare
technology by introducing an innovative approach to disease
prevention.
Methodology

 Data collection and preprocessing


 CNN Model Architecture
 Training Procedure
 Evaluation Metrics
 Testing on unknown data
Methodology
Data Collection and Preprocessing
 Dataset is collected.
 Eye_disease_classification dataset is used in this project which can be found on kaggle.
 Dataset are resized to same size while importing.
Prototype Development

CNN Model Architecture


 CNN generally contains 3 layers:
1. Convolutional layer
2. Pooling layer
3. Fully connected layer
Prototype Development
Training Procedure
 Dataset is split into training and testing data.
 Training data is used to train the data.
 Testing data is used to validate the model after training.
Prototype Development
Evaluation Metrics
 Confusion matrix is computed.
 Accuracy, precision, recall and F1 Score is calculated.
Prototype Development
Testing on Unknown data
 Prediction function is created.
 Testing the model with real data.
Literature Review
Author Title Findings

Indra Weni, Pradita Eko Prasetyo Detection of Cataract Based on Cataract Detection was
Utomo, Benedika Ferdian Image Features Using performed using CNN in different
Hutabarat, Muksin Alfalah Convolutional Neural Network epoch from 2 to 50 which
variable accuracy.

Q. Abbas Glaucoma-Deep: Detection of Glaucoma prediction was


Glaucoma Eye Disease on Retinal performed on Retinal Fundus
Fundus Images using Deep Image using CNN and deep-
Learning belief network.

Dr.E.Venkatesan, Dr.A.Naveen Study on Eye Cataract Disease Study of Cataract was performed
using Machine Learning based on factors like heredity,
Techniques birth defects, chronic disease,
excessive use of steroids and
bound eye injuries.
Thank You!

You might also like