Automatic glaucoma screening using optic nerve head measurements and random forest classifier on fundus images

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Science

Background:

• Glaucoma is a leading cause of irreversible blindness globally, characterized by optic nerve head damage.
• Early detection of glaucoma is crucial to prevent vision loss.
• Current diagnostic methods can be time-consuming and require expert interpretation.

Purpose of the Study:

• To develop and validate a novel algorithm for automated analysis of eye fundus images for glaucoma diagnosis.
• To measure key glaucoma indicators: Cup to Disc Ratio (CDR), ISNT rule, and Disc Damage Likelihood Scale (DDLS).
• To classify fundus images as either glaucoma or non-glaucoma using a random forest model.

Main Methods:

• An automated eye fundus image analysis algorithm was developed.
• The algorithm measures glaucoma indicators: CDR, ISNT rule, and DDLS.
• A random forest model was employed for binary classification of glaucoma cases.
• The method was validated on public (HRF) and local datasets.

Main Results:

• The algorithm achieved high performance metrics: sensitivity of 1, specificity of 0.93, and accuracy of 0.97.
• Achieved the highest classification accuracy compared to existing state-of-the-art methods.
• Demonstrated effective measurement of glaucoma-related indicators from fundus images.

Conclusions:

• The proposed algorithm offers a reliable and accurate method for automated glaucoma diagnosis.
• It can serve as a valuable computer-aided diagnosis tool for ophthalmologists in routine screening.
• This technique has the potential to improve early detection rates and patient outcomes for glaucoma.