Related Experiment Video
Updated: Jun 9, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
965
An efficient enhanced stacked auto encoder assisted optimized deep neural network for forecasting Dry Eye Disease
1Department of Electronics and Communication Engineering, Vins Christian College of Engineering, Chunkankadai, Nagercoil, Tamil Nadu, 629502, India. steffirajan7@gmail.com.
Scientific Reports
|October 22, 2024
Summary
This study presents a new deep learning model, ESAE-ODNN, for predicting Dry Eye Disease (DED). The model achieves high accuracy in early DED diagnosis by analyzing Meibomian Gland Dysfunction features.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Biomedical Engineering
Background:
- Meibomian Gland Dysfunction (MGD) and Dry Eye Disease (DED) are prevalent eye conditions affecting millions globally.
- Early detection and management of DED are crucial for improving patients' Quality of Life (QoL).
- Existing diagnostic methods may not fully capture the complex etiological factors of DED.
Purpose of the Study:
- To introduce an advanced deep learning model, the Enhanced Stacked Autoencoder-Optimised Deep Neural Network (ESAE-ODNN), for accurate and early prediction of DED.
- To enhance the identification and classification of DED by integrating sophisticated feature selection and extraction techniques.
- To improve the efficiency and accuracy of DED diagnosis through novel optimization strategies.
Main Methods:
- The ESAE-ODNN model utilizes feature selection (FS) incorporating chaotic maps and feature extraction (FE) via a stacked autoencoder (ESAE) to identify critical MGD-related features.
- A deep neural network (ODNN) classifier is employed for DED prediction, optimized using the Enhanced Quantum Bacterial Foraging Optimisation Algorithm (EQBFOA).
- The model incorporates SLSTM-STSA for enhanced classification accuracy and minimizes irrelevant/redundant features for robust performance.
Main Results:
- The proposed ESAE-ODNN model achieved a high classification accuracy of 96.34% in predicting DED.
- The method demonstrated efficiency in accurate identification, reduced computational complexity, and fine-tuned performance.
- Experimental evaluations confirmed the model's robustness in handling intricate features and high-dimensional data.
Conclusions:
- The ESAE-ODNN model offers a novel and efficient approach for the early diagnosis of DED.
- The integration of deep learning with advanced optimization techniques significantly improves the understanding and classification of MGD features.
- The proposed method outperforms existing state-of-the-art techniques in DED prediction, offering a promising tool for clinical application.
