Thursday 06 March 2025
A team of researchers has made a significant breakthrough in developing an AI-powered system that can accurately predict myopia, also known as nearsightedness, in children and adults using ultra-widefield fundus images. The new system, called CeViT, uses a combination of machine learning algorithms and copula theory to analyze the images and identify patterns associated with myopia.
Myopia is a common condition that affects millions of people worldwide, causing blurry vision at distances and increasing the risk of serious eye problems if left untreated. Early detection and treatment are crucial in preventing the progression of myopia, but current methods rely on subjective measurements and may not accurately predict the condition.
CeViT uses deep learning models to analyze ultra-widefield fundus images, which capture a wider field of view than traditional retinal scans. The system is trained on large datasets of images from both healthy eyes and those with myopia, allowing it to learn patterns and associations between image features and the presence of myopia.
The researchers used a combination of regression and classification tasks to evaluate CeViT’s performance. In their experiments, they found that CeViT outperformed existing methods in predicting axial length, a key indicator of myopia severity, with an accuracy rate of over 90%. The system also showed high accuracy in classifying images as either normal or showing signs of myopia.
One of the key advantages of CeViT is its ability to account for interocular asymmetry, where one eye may be more affected by myopia than the other. This is important because it allows the system to provide a more accurate prediction of an individual’s risk of developing myopia.
The potential benefits of CeViT are significant. Early detection and treatment of myopia could help prevent vision loss and reduce the burden on healthcare systems. The system could also be used to monitor patients with existing myopia, allowing for more effective management and prevention of complications.
While there is still much work to be done before CeViT can be used in clinical practice, this breakthrough has significant implications for the diagnosis and treatment of myopia. As researchers continue to refine the system, it may become an essential tool in the fight against this common and debilitating condition.
Cite this article: “AI-Powered System Accurately Predicts Myopia Using Ultra-Widefield Fundus Images”, The Science Archive, 2025.
Ai-Powered, Myopia, Nearsightedness, Ultra-Widefield Fundus Images, Machine Learning Algorithms, Copula Theory, Deep Learning Models, Axial Length, Interocular Asymmetry, Diagnostics
