Glaucoma disease diagnosis with an artificial algae-based deep learning algorithm.

Area of Science:

• Ophthalmology
• Computer Science
• Artificial Intelligence

Background:

• Glaucoma is an optic neuropathy causing blindness due to prolonged intraocular pressure.
• Deep learning aids disease diagnosis but traditional training methods are slow and prone to local minima.
• Efficient and accurate glaucoma diagnosis is crucial for preventing vision loss.

Purpose of the Study:

• To propose a novel deep learning-based decision support system for glaucoma diagnosis.
• To improve the efficiency and accuracy of deep learning model training for glaucoma detection.
• To overcome the limitations of traditional gradient-based learning in deep learning applications.

Main Methods:

• A two-stage system involving data preprocessing (normalization, mean absolute deviation) and deep learning training.
• Utilizing the artificial algae optimization algorithm for training the deep learning model.
• Comparing the proposed system against traditional gradient-based methods and other optimization algorithms (GA, PSO, BA, SSA, EO).

Main Results:

• The proposed system achieved superior performance in classification accuracy (0.9815), recall (0.9795), precision (0.9835), and F1-measure (0.9815).
• Demonstrated a significantly lower false positive rate (0.0165) compared to other methods.
• Outperformed state-of-the-art glaucoma detection algorithms in key performance metrics.

Conclusions:

• The novel deep learning system with artificial algae optimization offers a highly accurate and efficient approach to glaucoma diagnosis.
• This method addresses the limitations of traditional deep learning training, providing a promising tool for clinical application.
• The system's superior performance highlights its potential to aid ophthalmologists in early and precise glaucoma detection.