Artificial Intelligence Based Examination of the Field of Ophthalmology

• Published in: IEEE Uttar Pradesh Section International Conference on Electrical, Computer and Electronics (Online) Vol. 10; pp. 1617 - 1623
• Main Authors: Bhargavi, V.Sesha, Selvaraj, Krishnamoorthy, Koti, Valli Madhavi, Maheshwari, Rajeev Kumar, RadhaMahendran, S., Chandra, Sukumar
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2023
• Subjects: Artificial Intelligence; Classification algorithms; CNN (Convolutional Neural Networks); Convolutional neural networks; etc; Glaucoma; Ophthalmology; Reliability; Support vector machines; SVM (Support Vector Machines); Visualization
• ISSN: 2687-7767
• DOI: 10.1109/UPCON59197.2023.10434747

Early and correct diagnosis is critical for optimal care and prevention of visual loss due to glaucoma, the leading cause of permanent blindness globally. New possibilities for better glaucoma identification and monitoring have emerged the progress that has been made in the field of AI and machine learning. Using a unique use of both SVM and CNN, or support vector machines and convolutional neural networks, to solve a problem., this study automates the examination of ophthalmological results relevant to glaucoma. In this paper, we introduce a hybrid SVM-CNN algorithm that draws from the most promising aspects of both existing these two approaches to machine learning. In order to lay a solid groundwork for further investigation, SVM is combined first- class classification and feature extraction. CNN is used to refine and improve classification accuracy since it is a potent deep learning architecture that can automatically learn complicated characteristics from raw data. The algorithm is educated on a large collection of ophthalmic pictures used in glaucoma diagnosis, such as fundus photos, examinations of the eye's visual acuity and OCT scans (optical coherence tomography). Noise is removed, contrast is increased, and the format is standardized in these preprocessed photos. The presence or absence of glaucoma is subsequently determined by feature extraction and classification using the SVM-CNN hybrid model. This study's findings support the use of the SVM-CNN hybrid algorithm for reliable detection of glaucoma-related pathology in ophthalmic pictures. The suggested AI-based methodology has various benefits over conventional manual diagnosis approaches, including increased efficiency, more consistency, and the possibility of earlier detection. It can also help ophthalmologists by offering automated preliminary assessments, so the specialists can devote their time and energy where it is most needed. This study concludes the promise of AI-based approaches, and in particular the SVM-CNN hybrid algorithm, to revolutionize ophthalmology, particularly glaucoma diagnosis and treatment. We can improve patient outcomes and decrease the burden of this disabling eye illness by combining the benefits of SVM and CNN to increase the accuracy and efficiency of glaucoma identification. This paper highlights the significance of ongoing research and development on the promising subject of AI use in ophthalmology.