Scribd
Upload
5 views1 page

Explainable Semi-Supervised DR Detection

The document presents a project on an enhanced semi-supervised approach for diabetic retinopathy detection using contrastive learning, aimed at improving accuracy with limited labeled data. It integrates explainable AI techniques to provide interpretable insights into model predictions, facilitating clinical adoption. Experimental results indicate superior performance compared to traditional supervised methods, highlighting its potential for real-world healthcare applications.

Uploaded by

AKHIL Garena
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
5 views1 page

Explainable Semi-Supervised DR Detection

The document presents a project on an enhanced semi-supervised approach for diabetic retinopathy detection using contrastive learning, aimed at improving accuracy with limited labeled data. It integrates explainable AI techniques to provide interpretable insights into model predictions, facilitating clinical adoption. Experimental results indicate superior performance compared to traditional supervised methods, highlighting its potential for real-world healthcare applications.

Uploaded by

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

SRI VENKATESWARA COLLEGE OF ENGINEERING

(Autonomous)
DEPARTMENT OF DATA SCIENCE

Enhanced Diabetic Retinopathy Detection: An Explainable


Semi-Supervised Approach Using Contrastive Learning
Course: B. Tech Year: IV Sem: II Section: B Project ID: CSD26A02

ABSTRACT
Diabetic Retinopathy (DR) is a leading cause of vision impairment worldwide, requiring early and
accurate detection for effective treatment. Traditional supervised deep learning models demand large-
scale annotated datasets, which are often costly and time-consuming to obtain in the medical domain.
To address this challenge, we propose an enhanced semi-supervised framework for DR detection that
leverages contrastive learning to utilize both labeled and unlabeled retinal fundus images. Our
approach learns robust feature representations by maximizing inter-class separability while
preserving intra-class similarity, thereby improving classification performance with limited labeled
data. Furthermore, we integrate explainable AI techniques, such as Grad-CAM and attention maps, to
provide interpretable insights into model predictions, ensuring trust and transparency for clinical
adoption. Experimental evaluations on benchmark DR datasets demonstrate that the proposed method
achieves superior accuracy and generalization compared to conventional supervised baselines, while
requiring fewer annotations. This work highlights the potential of explainable semi-supervised
contrastive learning as a scalable and reliable solution for automated diabetic retinopathy detection in
real-world healthcare settings.

Project Team:

Student Roll Phone Signature of the


Name of the student
Number Number Number Student
S1 22BFA32117 SANA SAI PRANEETH 9515883403
S2 22BFA32112 POSA VAISHNAVI 7989808417
S3 22BFA32081 KOTHAPALLI KARTHIK 6304773886
S4 22BFA32085 M VENU GOPAL 9676501608

Name of the Guide: [Link].

Signature of the Guide Project Coordinator HOD/CSD

Academic Year 2025-2026

You might also like