Automatic Detection of Corneal ulcer by Machine Learning Algorithms

A corneal ulcer is a common ulcer in eyes which raised due to bacteria, over heat, pressure, viruses. Most of the time its not easily identified and leads to loss of vision. Medical practitioner needs more assistance from various test, laser treatments to sustenance corneal ulcer. Early detection may save the vision of corneal ulcer patients. This study utilized image processing with machine learning algorithms for corneal ulcer classification. This process starts from data augmentation, Feature extraction by GLCM and classification by Support Vector Machine (SVM), Random Forest(RF) and Decision Tree (DT). These algorithms experimented for corneal ulcer classification. Accuracy of all algorithms were compared with accuracy, F1 Score, Recall, Precision and found that SVM reached 95.67% of accuracy for corneal ulcer classification.