Deep Learning-Based Classification of Ocular Toxoplasmosis Fundus Images: A Comparative Study of CNN and SVM Models

Ocular toxoplasmosis poses a significant challenge in ophthalmology due to its potential to cause severe vision impairment. Current diagnostic methods rely heavily on subjective interpretation of fundus images, leading to inconsistencies and delays in treatment initiation. In this study, a novel approach for automated classification of ocular toxoplasmosis fundus images using state-of-the-art machine learning models is proposed. Leveraging the power of Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs), the methodology demonstrates promising results in accurately distinguishing between diseased and healthy fundus images. Through rigorous evaluation, our models exhibit high training and validation accuracies, with SVM classifiers trained on feature representations from VGG16, MobileNetV2, and InceptionV3 architectures achieving validation accuracies of 92.8%, 87.9%, and 93.9% respectively. The research opens avenues for early and reliable diagnosis of ocular toxoplasmosis, facilitating prompt treatment interventions and ultimately improving patient outcomes.