Ocular conditions may result in permanent eyesight loss. A proper diagnosis given early on can save someone’s vision because some symptoms manifest as soon as necessary. Since the human eye is susceptible to error, an automatic and exact method is required for this operation. The project’s goal is to develop a classical that can recognize an eye ailment from fundus photos. Iris recognition systems are among the most precise biometric technologies and have a huge potential for use in global security applications. This study investigated the impact of ocular pathology on iris recognition, specifically if ocular disease could lead to the failure of iris recognition systems. A prospective cohort of 54 patients with anterior segment eye illness who were seen at the acute referral unit of the Princess Alexandra Eye Pavilion in Edinburgh participated in the study. Before starting therapy and again at follow-up appointments after it had been administered, patients provided iris camera photos.The primary outcome metric was the mathematical difference between the iris recognition templates taken from patients' eyes before and after the eye condition therapy. Results revealed that the iris recognition templates taken from patients' eyes before and after the eye illness therapy differ mathematically. Results revealed that iris recognition performance was impressively resistant to most ocular illness states, including conjunctivitis, corneal oedema, and iridotomies (iris laser punctures). However, some patients with acute iris inflammation (iritis/anterior uveitis) experienced issues. The results demonstrate that Random Forest provides better prediction accuracy in less time than other ML approaches. The consequences of an individual having anterior uveitis could render obsolete the current recognition technologies. Iris recognition is a crucial biometric technology, so those who are creating and using it should be aware of any potential issues that could arise.
Paper
Full text
Ocular Disease Recognition using Machine Learning
Semantic Scholar · Medicine · 2023
Abstract
Ocular conditions may result in permanent eyesight loss. A proper diagnosis given early on can save someone’s vision because some symptoms manifest as soon as necessary. Since the human eye is susceptible to error, an automatic and exact method is required for this operation. The project’s goal is to develop a classical that can recognize an eye ailment from fundus photos. Iris recognition systems are among the most precise biometric technologies and have a huge potential for use in global security applications. This study investigated the impact of ocular pathology on iris recognition, specifically if ocular disease could lead to the failure of iris recognition systems. A prospective cohort of 54 patients with anterior segment eye illness who were seen at the acute referral unit of the Princess Alexandra Eye Pavilion in Edinburgh participated in the study. Before starting therapy and again at follow-up appointments after it had been administered, patients provided iris camera photos.The primary outcome metric was the mathematical difference between the iris recognition templates taken from patients' eyes before and after the eye condition therapy. Results revealed that the iris recognition templates taken from patients' eyes before and after the eye illness therapy differ mathematically. Results revealed that iris recognition performance was impressively resistant to most ocular illness states, including conjunctivitis, corneal oedema, and iridotomies (iris laser punctures). However, some patients with acute iris inflammation (iritis/anterior uveitis) experienced issues. The results demonstrate that Random Forest provides better prediction accuracy in less time than other ML approaches. The consequences of an individual having anterior uveitis could render obsolete the current recognition technologies. Iris recognition is a crucial biometric technology, so those who are creating and using it should be aware of any potential issues that could arise.