Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
June 19, 2026Discover Artificial IntelligenceOpen Access

Automated retinal disease classification from OCT images using particle swarm-optimized deep learning

View Full Paper
Ask AI
Bookmark
Share

Authors

AKAbdelaadim KhrissAEAissa Kerkour ElmiadMBMohammed Badaoui

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates improved accuracy in retinal disease classification using automated deep learning techniques, indicating significant efficiency benefits.

Key Points

  • The study aims to enhance retinal disease classification accuracy from OCT images through optimized hyperparameter tuning in CNNs.
  • Proposed Adaptive Multi-Strategy Particle Swarm Optimization (AMS-PSO) framework for tuning CNN hyperparameters.
  • Evaluated on a dataset of 84,495 OCT images across four pathology classes.
  • Achieved results using only 88 model evaluations compared to traditional grid search methods.
  • AMS-PSO achieved 95.24% test accuracy with 88 evaluations.
  • Outperformed Bayesian optimization (94.18%), grid search (92.87%), and standard PSO (93.87%).
  • Ablation studies confirmed the contributions of each optimization component.

Cite This Study

Khriss et al. (2026) studied this question.

synapsesocial.com/papers/6a34dde465a5b0777af2d6cahttps://doi.org/10.1007/s44163-026-01601-9
View Full Paper
Ask AI
Bookmark
Share