VISVESVARAYA TECHNOLOGICAL UNIVERSITY,

BELAGAVI, KARNATAKA

A Project Synopsis On

“AUTOMATED DETECTION OF RETINOPATHY USING QUANTUM MACHINE

LEARING AND DEEP LEARNING TECHNIQUES”
Submitted for partial fulfillment of the requirement for the award of the Bachelor degree in Computer
Science and Engineering during the year 2024-2025.

Submitted by

ANUSHA N 4CA21CS010

CHAITHANYA H D 4CA21CS011

LAVANYA M S 4CA21CS042

NAMRATHA R GOWDA 4CA21CS054

Under the Guidance of

Prof. RAKSHITHA B H
CIT COLLEGE, MANDYA

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

2024-25
 CAVERY INSTITUTE OF TECHNOLOGY MANDYA
SIDDAIAHNAKOPPLU GATE, SUNDAHALLI,
KARNATAKA- 571401 ,2024-25

ABSTRACT.

There are several deep learning techniques that are used to perform the predictive analytics over
big data in various medical tasks. Predictive analytics in medical healthcare is a challenging task
yet ultimately helping the practitioners handle big data-informed timely decisions about patient’s
medical health and treatment. This paper discusses the predictive analytics in healthcare. Patient’s
medical record is obtained for experimental research. The two architectures of deep learning are
implemented. Performance and accuracy of these applied algorithms are implemented and
compared. Different deep learning techniques used in this research that reveals which algorithm is
best suited for the prediction of diabetes over the patient. This paper aims to help doctors and
practitioners in early stage to predict diabetic retinopathy using deep learning techniques.
Keywords: Dataset, Res Net, Convolutional Neural Network.
 INTRODUCTION
Healthcare industry is a very large and sensitive meta data and must be carefully handled. One of
the growing extremely fatal diseases all over the world is Diabetes Mellitus. Medical professionals
require a reliable prediction system to diagnose this disease. Some of the useful deep learning
techniques for examining the data from diverse perspectives and synopsizing it into valuable
information. The accessibility and availability of huge amounts of data are able to provide us useful
knowledge unless certain data mining techniques are applied to it. The main goal to this is to
determine new patterns and interpret these patterns to deliver acute significant and useful
information for the process. Diabetes may leads to heart disease, kidney disease, nerve damage,
and blindness. Diabetes data mining in an efficient way for a crucial concerns. The data mining
techniques and its way be discovered to create the appropriate approaches and techniques for an
efficient classification of Diabetes dataset and in extracting it data patterns. In this study, medical
bioinformatics analyses is done in diabetes prediction. The WEKA software is employed as a
mining tool in diabetics diagnosis. The Pima Indian diabetes Data base was acquired from UCI
repository that used for future analysis. The dataset was researched and analyzed to build an
effective model that used to predict and diagnoses diabetes disease. Diabetic Retinopathy may be
a complication of diabetes that's caused thanks to the changes within the blood vessels of the retina
and is one among the leading causes of blindness in the developed world. Up to this, Diabetic
Retinopathy remains screened manually by ophthalmologist which may be a time consuming
process and hence this paper aims at automatic diagnosis of the disease into its different stages
using deep learning. In our approach, we trained a Deep Convolutional Neural Network model on
an outsized dataset consisting of around 35,000 images to automatically diagnose and thereby
classify high resolution fundus images of the retina into five stages supported their severity. Within
this paper, an application system is made which takes the input parameters because the patient’s
details alongside the fundus image of the attention. A trained deep Convolutional neural network
model will further extract the features of the fundus images and later with the assistance of the
activation functions like Relu and SoftMax along with optimizing techniques like Adam an output
is obtained.
The output obtained from the Convolutional Neural Network (CNN) model and therefore the
patient details will collectively make a uniform report. In this study, we aim to apply the
bootstrapping resembling technique to enhance the accuracy and then applying ResNet, CNN and
compare their performance. The idea of Machine learning is to predict the longer term from past
data. Machine learning focuses on the event of Computer Programs which will change when
exposed to new data and therefore the basics of Machine Learning, implementation of an easy
machine learning algorithm using python. It feed the training data to an algorithm, and therefore
the algorithm uses this training data to offer predictions on a replacement test data.
Machine learning are often roughly separated in to three categories. Supervised learning program
is given to both the input file and therefore the corresponding labelling to find out the info that has
got to be labelled by a person's being beforehand. Unsupervised learning has no labels. It is directly
fed to the learning algorithm. This algorithm has got to discover the clustering of the input file.
Finally, Reinforcement learning dynamically interacts with its environment and it receives positive
or feedback to enhance its performance. Deep learning is a part of machine learning in artificial
intelligence that has networks able to adapt in learning unsupervised data that are unlabeled or
unstructured. This process is otherwise known as deep neural learning or deep neural networks.
The Deep Learning consists of an algorithm called Convolutional Neural Network
(ConvNet/CNN) which gets the input image, assign importance (learnable weights and biases) to
various aspects/objects in that image and be able to distinguish one from the other. The pre-
processing required in a ConvNet is much lesser as compared to the other classification algorithms.
In Deep Learning primitive methods filters are hand-engineered that comes with enough training,
Convolutional Networks can learn these filters/characteristics.
Diabetic retinopathy is a medical complication that is caused by the damage to the blood vessels
of the light-sensitive tissue which is present at the back of the eye, retina, which can gradually lead
to complete blindness and various other eye problems depending on the severity of Diabetic
Retinopathy. It is observed that 40% − 45% of diabetic patients are likely to have DR in their life,
but due to lack of knowledge and delayed diagnosis, the condition escalates quickly. Diabetes was
once thought of as a disease of the affluent but it's now reached epidemic proportion in both
developed and developing countries. Currently, a minimum of 366 million people worldwide has
diabetes, and this number is probably going to extend because of an aging global population
Globally, the quantity of individuals with DR will grow from 126.6 million in 2010 to 191.0 million
by 2030, and that we estimate that the quantity with vision-threatening diabetic retinopathy
(VTDR) will increase from 37.3 million to 56.3 million, if prompt action is not taken. Diabetes is
the best disease to apply deep learning principles too.
 LITERATURE SURVEY
Sl. Authors Year of Title of the paper Methodology Limitations
No publica
tion
1 Jing Wang, Liu 2020 Multi-Label The purpose of The amount
Yang, Zhanqiang Classification of Fundus this study is to of data for some diseases is
Huo , Weifeng Images With establish a very limited, which makes it
He, and Junwei EfficientNet. framework for very difficult to improve the
performance of a network.
Luo automatic
Another basic limitation comes
identification from the black box of the
of multi-label nature of deep networks. The
fundus network automatically learns
diseases, and features from images, but the
to achieve it by specific features learned are
designing a unknown.
corresponding
ensemble
model.
2. Balla Goutam 2022 A Comprehensive The proposed The models
, Mohammad Review of Deep review focuses trained on IDRiD,
Farukh Hashmi Learning mainly on Messidor, DRIVE datasets may
, (Senior Member, Strategies in Retinal providing an in not perform well on other
datasets. These may not be
Ieee), Disease Diagnosis depth review
suitable for efficient model
Zong Woo Geem , Using Fundus Images of various DL training,
(Senior Member, strategies
Ieee), and Neeraj recently
Dhanraj Bokde implemented
for retinal
disease
diagnosis
using fundus
images. This
study
also intends to
outline
possible future
directions for
new
researchers
interested in
AI-based
retinal disease
diagnosis.
3. Juan Carrillo 2019 Glaucoma Detection This work The vessels segmentation
, Lola Bautista, Using Fundus Images of presents a requires an
Jorge Villamizar, The Eye computational improvement due to some fails
Juan Rueda, Mary tool for in different images and residual
Sanchez and automatic noise after the segmentation.
Daniela Rueda glaucoma
detection from
fundus images
of the eye. This
work propose a
novel method
for cup
segmentation,
which shows
an
improvement
in the accuracy
compared to
other methods
4. Ayesha Kazi, 2018 Processing Retinal This paper It cannot determine the
Prerna Sukhija, Images to Discover aims to not presence of a variation in the
Miloy Ajmera, Diseases only image
Kailas Devadkar accurately (like highlighting, tessellation
etc) preceding the
classify the
classification step, for training
disease the neural network, and to
into one of the use unique combinations for
three possible different cases.
abnormalities
or assert that
it is a healthy
retina. The
image may fall
into more than
one category.
For example, a
retina which
shows any
presence of
Diabetic
Retinopathy
could also
show hints of
Glaucoma
and/or
Cataract
 PROBLEM STATEMENT
The problem addressed in this study is the need for a cost-effective and scalable solution for the
detection of diabetic retinopathy (DR), a leading cause of blindness in adults with diabetes. Current
methods for DR detection are time-consuming, expensive, and require specialized equipment and
trained personnel, making them inaccessible in many regions. The objective of this study is to
develop and evaluate a deep learning model for automated DR detection from retinal images, which
can improve the accessibility and affordability of DR screening and diagnosis. • The current
screening methods for diabetic retinopathy (DR) are inefficient, leading to delayed diagnosis and
poor disease management. Traditional screening techniques require manual examination by trained
professionals, which can be time-consuming, costly, and prone to human error. This can result in
missed diagnoses and delayed treatment, leading to irreversible damage to the retina and
potentially permanent vision loss. • Moreover, as the prevalence of diabetes continues to rise
globally, the burden on healthcare systems for DR screening is becoming increasingly significant.
Therefore, there is a pressing need for an automated and efficient DR screening method to improve
early detection and management of the condition. In recent years, deep learning techniques have
shown promise in aiding the diagnosis of DR, and this study aims to investigate their potential for
automated

EXISTING SYSTEM

Apart from a binocular model for the various classes of Diabetic Retinopathy detection task is also
trained and evaluated to further prove the effectiveness of the binocular design. The final result
shows that, on a 10% validation set, the binocular model achieves a kappa score of 0.829 which is
higher than that of existing non ensemble model. In the end the analogy between confusion
matrices obtained through models with paired and unpaired inputs is performed and it demonstrates
that the binocular architecture does improve the classification performance.
 PROPOSED SYSTEM

The proposed system for the automated detection of Retinopathy of Prematurity (ROP) utilizes a
hybrid approach integrating Deep Learning techniques to enhance diagnostic accuracy and speed.
The system employs deep convolutional neural networks (CNNs) for feature extraction from
retinal images, leveraging QML algorithms to optimize model parameters and improve
classification performance. By utilizing quantum-enhanced learning, the model can process
complex patterns in high-dimensional medical data more efficiently. The system aims to assist
ophthalmologists by providing a reliable, automated ROP screening tool that minimizes the need
for manual interpretation, reducing human error and improving early intervention outcomes

MOTIVATION
The motivation behind developing a deep learning model for diabetic retinopathy detection is the
significant burden of this disease on patients, healthcare systems, and society. Early detection and
treatment of DR are critical to preventing blindness and improving patient outcomes. However,
current methods for DR detection are expensive and inaccessible in many regions, leading to delays
in diagnosis and treatment. By developing a cost-effective and scalable solution for DR detection
using deep learning, this study aims to improve the accessibility and affordability of DR screening
and diagnosis, ultimately improving patient outcomes and reducing the burden on healthcare
systems.
 SCOPE

The scope of this study is to develop and evaluate a deep learning model for automated diabetic
retinopathy (DR) detection from retinal images. The study will use publicly available retinal image
datasets with labeled DR severity to train and validate the deep learning model. The deep learning
model will use a combination of convolutional neural networks (CNNs) and recurrent neural
networks (RNNs) for feature extraction, sequence modeling, and classification of DR severity. The
performance of the deep learning model will be compared with other state-of-the-art DR detection
methods to assess its superiority and effectiveness. The developed deep learning model can have
significant clinical implications, including early detection and treatment of DR, reducing the
workload of ophthalmologists, and improving patient outcomes. However, the study may face
limitations such as limited sample size and biased data, which can affect the generalizability of the
deep learning model. The scope of this study is limited to the development and evaluation of a
deep learning model for automated DR detection and does not include implementation and
validation in real-world clinical settings.
 OBJECTIVES
The objectives of this study are

1. To develop a deep learning model for automated diabetic retinopathy (DR) detection from retinal
images.
2. To evaluate the performance of the deep learning model using standard techniques, such as
backpropagation and gradient descent, and performance metrics such as accuracy, sensitivity, and
specificity.
3. To compare the performance of the deep learning model with other state-of-the-art DR detection
methods to assess its superiority and effectiveness.
4. To assess the clinical implications of the developed deep learning model, including early
detection and treatment of DR, reducing the workload of ophthalmologists, and improving patient
outcomes.
5. To identify any limitations or challenges faced in the development and evaluation of the deep
learning model, such as limited sample size or biased data, that may affect the generalizability of
the model.
6. To provide recommendations for future research in the field of DR detection using deep learning,
such as the implementation and validation of the model in real-world clinical settings.
 SYSTEM REQUIREMENTS

Hardware Requirements
• System: Pentium IV 2.4 GHz/intel i3/i4.
• Hard Disk: 40GB.
• Monitor: 15 VGA Color.
• RAM: 512 Mb Minimum

Software Requirements
• Operating system: Windows XP/ Windows10
• Software Packages: Ten Sor flow, OpenCV
• Coding Language: Python.
 METHODOLOGY

1. Data Collection:
• Gather a diverse and extensive dataset of retinal images, including images from diabetic patients
with varying stages of retinopathy and images from healthy individuals for comparison.
2. Preprocessing:
• Preprocess the retinal images to enhance their quality and prepare them for analysis. This may
involve resizing, contrast adjustment, noise reduction, and image normalization.
3. Data Splitting:
• Divide the dataset into training, validation, and testing sets. Cross-validation techniques may also
be applied to ensure robust model evaluation.
4. Feature Extraction:
• Utilize image processing techniques and deep learning methods to extract relevant features from
retinal images, such as blood vessel patterns, hemorrhages, exudates, and microaneurysms.
5. Deep Learning Model Selection:
• Choose an appropriate deep learning architecture for image classification. Convolutional Neural
Networks (CNNs) are commonly used due to their effectiveness in image analysis.
6. Model Training:
• Train the selected deep learning model on the training dataset. This involves adjusting model
parameters to learn the patterns and characteristics associated with diabetic retinopathy.
7. Hyperparameter Tuning:
• Optimize hyperparameters, including learning rate, batch size, and architecture-specific
parameters, to improve model performance.
8. Validation and Evaluation:
• Validate the model's performance on the validation dataset and make necessary adjustments.
Employ evaluation metrics like accuracy, precision, recall, F1-score, and ROC-AUC to assess
model performance.
9. Testing:
• Assess the model's performance on the separate testing dataset to ensure it generalizes well to
unseen data.
 ADVANTAGES

1. Improved Accuracy:
Quantum machine learning leverages quantum computing principles to process information in
ways classical computers cannot. This can lead to more nuanced models that recognize complex
patterns in retinal images, enhancing diagnostic accuracy. By training on large datasets, these
models learn to identify ROP stages with high precision.
2. Rapid Processing:
Quantum algorithms can perform calculations at exponentially faster rates than classical
algorithms. This speed enables the analysis of vast quantities of imaging data quickly, which is
critical in neonatal care, where time-sensitive decisions can prevent vision loss.
3. Advanced Feature Extraction:
Deep learning techniques, particularly convolutional neural networks (CNNs), excel at image
processing. They automatically extract relevant features from images, such as abnormal blood
vessel growth, which are crucial for identifying ROP. This capability allows for more thorough
evaluations compared to traditional methods.
4. Scalability:
The automation of ROP detection can be applied in various healthcare settings, from specialized
hospitals to remote clinics. By deploying these systems widely, healthcare providers can screen
larger populations, ensuring more premature infants receive necessary evaluations.
5. Real-Time Monitoring:
Automated systems can continuously analyze patient data, enabling healthcare providers to
monitor infants in real time. If a significant change in retinal condition is detected, immediate
intervention can be initiated, improving overall care and outcomes.
6. Reduced Interobserver Variability:
Human assessments can vary significantly due to individual expertise and subjective interpretation.
Automated detection systems provide consistent evaluations, reducing discrepancies and
enhancing trust in the results.
7. Integration of Multimodal Data:
These advanced techniques can incorporate various forms of data, such as clinical histories,
imaging studies, and demographic information. This holistic approach allows for a more
comprehensive risk assessment and better-informed clinical decisions.
8. Predictive Modeling:
Using historical data, quantum and deep learning models can identify risk factors associated with
ROP progression. By predicting which infants are at higher risk, healthcare providers can prioritize
screenings and tailor preventive measures accordingly.

9. Cost-Effectiveness:
Automating ROP detection can significantly reduce the costs associated with manual screenings
and consultations. By streamlining processes and improving early detection, healthcare systems
can allocate resources more efficiently.

10. Continuous Improvement through Learning:

Deep learning models can continuously improve as they are exposed to more data over time. This
adaptive learning means that the systems can evolve to incorporate the latest research and clinical
insights, further enhancing their diagnostic capabilities.
 APPLICATIONS

1. Overview of ROP:
- ROP is a potentially blinding condition in premature infants, requiring early detection and
treatment.
- Traditional screening methods involve manual examination by trained ophthalmologists, which
can be resource-intensive and subject to variability.
2. Deep Learning Techniques:
- Convolutional Neural Networks (CNNs): Used for image analysis, CNNs can be trained on retinal
images to detect features indicative of ROP, such as vascular changes and abnormalities.
- Transfer Learning: Utilizing pre-trained models can improve accuracy and reduce the amount of
labeled data required, which is valuable given the limited dataset for ROP.
3. Quantum Machine Learning Techniques:
- Quantum-enhanced Feature Extraction: QML can potentially identify complex patterns in retinal
images more efficiently than classical methods. Algorithms like Quantum Support Vector
Machines or Quantum Neural Networks can be leveraged.
- Speed and Efficiency: Quantum algorithms can process data at unprecedented speeds, allowing
for real-time analysis and quicker diagnosis.
4. Integration of QML and Deep Learning:
- Hybrid Models: Combining classical deep learning models with quantum algorithms can enhance
performance, particularly in handling large datasets and improving generalization.
- Data Encoding: Techniques for encoding classical data into quantum states can be developed to
maximize the advantages of quantum processing.
5. Clinical Implementation:
- Automated Screening Tools: The integration of these technologies can lead to the development
of automated screening systems that can operate in neonatal units, providing rapid assessments and
prioritizing cases for specialist review.
- Telemedicine Solutions: With automated tools, ROP screenings could be conducted remotely,
enabling access to specialist care in underserved regions.
6. Challenges and Considerations:
- Data Availability: Large, annotated datasets are essential for training models effectively.
- Interpretability: Ensuring that automated systems provide interpretable results for clinicians is
crucial for trust and adoption.
- Regulatory Approval: Any system developed must meet clinical guidelines and regulatory
standards for medical devices.

7. Future Directions:
- Research should focus on improving model accuracy, reducing computational overhead, and
developing robust clinical validation studies.
- Exploring more advanced quantum algorithms could further enhance detection capabilities.
 FLOW CHART

Fig.1: Flowchart of the Project

 REFERENCES
1. Abràmoff, M. D., Lou, Y., Erginay, A., Clarida, W., Amelon, R., Folk, J. C., & Niemeijer, M. (2016).
Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of
deep learning. Investigative Ophthalmology & Visual Science, 57(13), 5200-5206.
2. Bhaskaranand, M., Ramachandra, C., Bhat, S., & Cuadros, J. (2020). Automated diabetic retinopathy
screening and monitoring using deep learning. Journal of Healthcare Engineering, 2020, 1-13.
3. Burlina, P. M., Joshi, N., Pekala, M., Pacheco, K. D., Freund, D. E., Bressler, N. M., & Wong, T. Y. (2017).
Automated grading of age-related macular degeneration from color fundus images using deep convolutional
neural networks. JAMA ophthalmology, 135(11), 1170-1176.
4. Gulshan, V., Peng, L., Coram, M., Stumpe, M. C., Wu, D., Narayanaswamy, A., ... & Webster, D. R. (2016).
Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal
fundus photographs. Jama, 316(22), 2402-2410.
5. Kermany, D. S., Goldbaum, M., Cai, W., Valentim, C. C., Liang, H., Baxter, S. L., ... & Keane, P. A. (2018).
Identifying medical diagnoses and treatable diseases by image-based deep learning. Cell, 172(5), 1122-
1131.
6. Li, Z., Keel, S., Liu, C., He, Y., Meng, W., Scheetz, J., ... & Ting, D. S. (2020). An automated grading
system for detection of vision-threatening referable diabetic retinopathy based on color fundus photographs.
Diabetes care, 43(8), 1748-1755.
7. Osareh, A., Shadgar, B., & Markham, R. (2012). A survey of computer-aided diagnosis of ocular diseases.
Computer Methods and Programs in Biomedicine, 108(1), 407-433.
8. Rajalakshmi, R., Subashini, R., Anjana, R. M., Mohan, V., & Deepa, M. (2018). Automated diabetic
retinopathy detection in smartphone-based fundus photography using artificial intelligence. Eye, 32(6),
1138-1144.