Published Paper Details:

Amblyopia, commonly known as "lazy eye," is a neurodevelopmental disorder characterized by impaired vision, often due to misalignment, refractive issues, or other visual deficits. Early detection is critical to prevent long-term vision impairment. This project proposes a dual model approach leveraging artificial intelligence to enhance amblyopia diagnosis through both image-based and genetic risk analysis. The first model utilizes an image classifier to analyze eye images, identifying physical markers of amblyopia using convolutional neural networks (CNNs) or Vision Transformers. The second model incorporates Random Forest and Naive Bayes algorithms to evaluate genetic predispositions and familial risk factors, offering a probabilistic assessment of a child's susceptibility to developing amblyopia. A hybrid model integrates both the image-based and genetic-risk predictions, producing a unified risk score to improve diagnostic accuracy. This comprehensive diagnostic framework holds potential for early detection of amblyopia and a more personalized, predictive approach to pediatric eye care. By combining visual and genetic data, the proposed method aims to provide clinicians with a robust tool to address amblyopia risk more accurately and efficiently.