Computer-Aided Diagnosis of Anterior Segment Eye Abnormalities using Visible Wavelength Image Analysis Based Machine...

Area of Science:

• Ophthalmology
• Computer Science
• Medical Imaging

Background:

• Eye diseases, particularly cataracts and corneal arcus, significantly impact the elderly population.
• Early detection and grading of anterior segment eye abnormalities are crucial for effective ophthalmological care.
• Computer-aided diagnosis (CAD) systems offer potential for mass screening and objective assessment.

Purpose of the Study:

• To develop and evaluate a multiclass computer-aided diagnosis (CAD) system for anterior segment eye abnormalities.
• To utilize visible wavelength (VW) eye images for automated diagnosis.
• To assess the system's accuracy, sensitivity, and specificity for clinical application.

Main Methods:

• A CAD system was developed using visible wavelength (VW) eye images.
• Image pre-processing included specular reflection removal.
• Iris circle segmentation was performed using a circular Hough Transform (CHT).
• Feature extraction involved first-order statistics and wavelet-based features.
• Classification was conducted using a Support Vector Machine (SVM) with the Sequential Minimal Optimization (SMO) algorithm.

Main Results:

• The system was tested on 228 VW eye images across three classes of anterior segment eye abnormalities.
• The proposed method achieved a high predictive accuracy of 96.96%.
• The system demonstrated excellent performance with 97% sensitivity and 99% specificity.

Conclusions:

• The developed computer-aided diagnosis system shows significant potential for clinical applications in ophthalmology.
• The method provides an accurate and efficient approach for screening and grading anterior segment eye abnormalities.
• The use of VW images and advanced feature extraction/classification techniques contributes to the system's high diagnostic performance.