Development of a Computer Assisted System for Automatic Glaucoma Detection Using Pigeon Inspired Optimization

Glaucoma is one of the eye conditions that is considered the main cause for significant irreversible loss of vision if the retina is not diagnosed and treated in the early stages. However, glaucoma usually does not give any clear symptoms during its initial stages. Although SVM, KNN and Random Forest, among other machine learning techniques, have been optimally used for the detection of glaucoma, these techniques tend to struggle with attaining accuracy. This study proposes a cutting-edge and novel approach which employs Pigeon Inspired Optimization (PIO) to determine the best threshold for the classification of glaucoma. By adjusting the threshold value with respect to essential retinal characteristics like microaneurysms, exudates and hemorrhages, the method suggests the possibility of bettering the glaucoma diagnostic accuracy. The more reliable decision making due to optimization processes offers a contemporary solution for the accuracy and reliability of diagnosis in its pronominal stages.