Data driven based correlation and prediction analysis of ocular surface diseases
Aims: The overall goal of this project is to analyze the crucial determinants and assess risk factors causally related to different kinds of ocular surface diseases, and to explore factors affecting corneal biomechanics. In addition to that, a deep learning system to predict the probability of occurrence and the corneal biomechanical alterations in these diseases will be built.
Ocular surface diseases, especially keratopathy and dry eye, are very serious challenges that bring enormous burden to patients and societies. Because no significant symptom can be observed at the early stages, many diseases may not be discernible until advanced stage, resulting in the impairment in visual functioning and vision-related quality of life. Along with the rapid development of deep learning these years, many deep learning models have been established in many ocular diseases. However, these models most focus on fundus lesions, we wonder whether ocular surface diseases could be predicted by using risk factors. Therefore, we want to use these data from different dimensions to explore the risk factors causally related to diseases, and finally build deep learning models to predict corneal biomechanical alterations and ocular surface diseases. Also, we need to validate these models.
If these models are successfully developed, many bothered by these diseases or at high risk of developing ocular surface lesions will fundamentally benefit from it, reducing medical cost.