EUVIMEDEUVIMED entdecken
Crossref

Deep Learning based: Diabetic Retinopathy Detection using Retinal Fundus Images

Shreeharsh Shinde Shreeharsh Shinde, Rushikesh Ranaware Rushikesh Ranaware, Tushar Farande Tushar Farande, Vihang Dandawar Vihang Dandawar, Y. N. Sakhare Y. N. Sakhare

International Journal of Creative and Open Research in Engineering and Management · 18.06.2026 · 02 (06) · 1-6

DOI: 10.55041/ijcope.v2i6.248 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Creative and Open Research in Engineering and Management, online erschienen am 18.06.2026. Verfasst von Shreeharsh Shinde Shreeharsh Shinde und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Diabetic Retinopathy (DR) remains a leading cause of vision impairment globally, necessitating robust automated screening tools to alleviate the burden on ophthalmologists. This paper presents an end-to-end production-grade pipeline for DR grading using a fine-tuned MobileNetV2 architecture. Leveraging a unified dataset of 1.15 lakh retinal fundus images from EyePACS, APTOS 2019, and Messidor-2 repositories, we address the critical challenges of data noise and class imbalance. Our methodology integrates a sophisticated OpenCV-based pre-processing engine utilizing Contrast Limited Adaptive Histogram Equalization (CLAHE) and smart contour cropping to amplify microscopic lesions. To ensure viability for edge deployment on constrained hardware (2GB VRAM), we implemented selective fine-tuning with Batch Normalization freezing and utilized Cate-gorical Focal Loss to prioritize high-severity cases. Experimental results demonstrate a peak validation accuracy of 75.83% and a high clinical specificity, achieving a 94% recall for healthy retinas. The system is deployed via a full-stack architecture featuring a FastAPI backend and a Streamlit dashboard, incorporating Test-Time Augmentation (TTA) to provide stable, clinical-grade diagnostics. Index Terms—Deep Learning, Skin Disease Classification, Computer-Aided Diagnosis, EfficientNetV2, Vision Transformer (ViT), Image Segmentation, Medical Image Analysis

Zitieren

APA

Shinde, S. S. S., Ranaware, R. R. R., Farande, T. F. T., Dandawar, V. D. V., & Sakhare, Y. N. S. Y. N. (2026). Deep Learning based: Diabetic Retinopathy Detection using Retinal Fundus Images. International Journal of Creative and Open Research in Engineering and Management, 02(06), 1-6. https://doi.org/10.55041/ijcope.v2i6.248

Vancouver

Shinde SSS, Ranaware RRR, Farande TFT, Dandawar VDV, Sakhare YNSYN. Deep Learning based: Diabetic Retinopathy Detection using Retinal Fundus Images. International Journal of Creative and Open Research in Engineering and Management. 2026;02(06):1-6. doi:10.55041/ijcope.v2i6.248

BibTeX
@article{Shinde2026,
  author = {Shinde, Shreeharsh Shinde Shreeharsh and Ranaware, Rushikesh Ranaware Rushikesh and Farande, Tushar Farande Tushar and Dandawar, Vihang Dandawar Vihang and Sakhare, Y. N. Sakhare Y. N.},
  title = {Deep Learning based: Diabetic Retinopathy Detection using Retinal Fundus Images},
  journal = {International Journal of Creative and Open Research in Engineering and Management},
  year = {2026},
  volume = {02},
  number = {06},
  pages = {1--6},
  doi = {10.55041/ijcope.v2i6.248}
}
Kostenlos starten →