Related Experiment Videos
Neural networks to identify glaucoma with structural and functional measurements
L Brigatti1, D Hoffman, J Caprioli
1Department of Ophthalmology and Visual Science, Yale University School of Medicine, New Haven, CT 06520, USA.
American Journal of Ophthalmology
|May 1, 1996
Summary
Neural networks effectively identify early glaucoma by analyzing optic nerve and visual field data. This technology aids in classifying glaucomatous eyes with high accuracy, improving early diagnosis.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Medical Diagnostics
Background:
- Glaucoma is a leading cause of irreversible blindness.
- Early detection of glaucomatous damage is crucial for preserving vision.
- Current diagnostic methods can be limited in identifying early-stage disease.
Purpose of the Study:
- To evaluate the efficacy of neural networks in distinguishing between normal and glaucomatous eyes.
- To assess the performance of neural networks using both structural and functional ocular measurements.
Main Methods:
- Utilized a database of 185 glaucomatous eyes and 54 normal control eyes.
- Employed various neural network algorithms, including backpropagation networks.
- Input data comprised automated visual field indices and structural parameters from computerized image analysis.
Main Results:
- A backpropagation network correctly identified 88% of all eyes.
- Achieved 90% sensitivity and 84% specificity when using combined structural and functional data.
- Neural networks trained on structural or functional data alone showed slightly lower, but still significant, identification rates.
Conclusions:
- Neural networks can accurately identify early glaucomatous damage.
- Analysis of optic nerve and visual field variables by neural networks aids in probabilistic assessment of glaucoma presence.
- This approach holds promise for improving early glaucoma diagnosis and patient management.