Approved Research
Predicting myocardial infarction through retinal scans
Approved Research ID: 96446
Approval date: March 2nd 2023
Lay summary
Predicting myocardial infarction through retinal scans. The project will include ML and DL models, Embedded system and flutter.
Our aim is to make the project contribute to protecting many patients from having a heart attack, as well as saving the costs of repeated examinations.
Through the examination, we can find that slight changes in the small blood vessels in the retina, or the back of the eye, can show signs that a person may have a heart attack, the role of the AI model is to monitor these slight changes and this technique increases the efficiency rate by ten percent more than using demographic data only.
The model is made using two types of data, the first is data for retinal examination of more than sixty thousand people, and the other is data that combines retinal examination and CMR for five thousand people. The smaller sample is used to train and test the data and then used as a template to train the larger data.
Project will last for at least 10 months from October 2022 and It will have a medical effect in protecting the lives of many blood pressure and heart patients