Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 25, 2026American Journal of Artificial IntelligenceOpen Access

VGG-19 Transfer Learning Technique for Automated Multi-Class Retinal Disease Detection: Model Development and Validation on a Ghanaian Fundus Image Dataset

View Full Paper
Ask AI
Bookmark
Share

Authors

MAMichael Adusei-NsowahFNFred Adusei NsowahSASamuel Afari

Discussion

Loading...

Member takes

Overview

Demonstrates automated retinal disease detection in a Ghanaian population, highlighting potential for early diagnosis and telemedicine.

Key Points

  • The aim is to develop an automated system for detecting various retinal diseases using deep learning techniques.
  • Utilized a VGG-19 convolutional neural network architecture for image analysis
  • Employed augmentation techniques on 184 retinal fundus images from Ghana
  • Conducted a two-stage classification to identify healthy vs unhealthy images
  • Classified unhealthy images into specific retinal conditions including glaucoma and diabetic retinopathy
  • Evaluated performance using metrics such as accuracy, precision, recall, and AUC
  • Achieved 97.31% accuracy in detecting retinal diseases
  • Precision reached 96.85% and recall at 98.06%
  • AUC of 0.993 indicates high classification ability
  • Demonstrated effectiveness for early diagnosis and automated screening of retinal conditions

Cite This Study

Adusei-Nsowah et al. (2026) studied this question.

synapsesocial.com/papers/69c37b74b34aaaeb1a67dcfehttps://doi.org/10.11648/j.ajai.20261001.22
View Full Paper
Ask AI
Bookmark
Share