Objective assessment of corneal staining using digital image analysis.

Area of Science:

• Ophthalmology
• Medical Imaging
• Computational Biology

Background:

• Corneal staining is a key indicator of ocular surface disease.
• Current grading methods are subjective and can vary between clinicians.
• Objective and reproducible assessment of corneal staining is needed.

Purpose of the Study:

• To validate a novel objective digital image analysis technique for evaluating corneal staining.
• To compare the new technique's performance against established clinical grading scales.

Main Methods:

• A new algorithm combining Difference of Gaussians (DoG) and RGB/HSV color models was developed.
• Image enhancement techniques including median filter, Otsu thresholding, and CLAHE were employed.
• Photographs were graded by clinicians using Oxford and National Eye Institute/Industry (NEI) guidelines for validation.

Main Results:

• The algorithm demonstrated strong correlation with the Oxford scheme (R=0.850) and NEI guidelines (R=0.903).
• Excellent repeatability was observed for the objective measurements (R=0.994).
• The technique proved effective across various ocular surface diseases.

Conclusions:

• The developed digital image analysis technique provides an objective and reliable method for corneal staining assessment.
• This algorithm shows excellent agreement with subjective clinical grading, potentially improving diagnostic consistency.
• The technique is suitable for objective evaluation of corneal staining, irrespective of the underlying disease.