Retinopathy of prematurity (ROP) is a major cause of childhood blindness worldwide and even for a subject area specialist, making a clinical diagnosis is both subjective and variable.
A deep learning algorithm was implemented and validated after training with a data set of over 5000 retinal photographs. Results show a 91% accuracy of the algorithm, compared with 82% average accuracy rate among clinicians with expertise in that area.
The results hold potential for disease detection and monitoring among infants at risk for retinopathy of prematurity. The famous musician Stevie Wonder is known to have suffered from the debilitating illness that resulted in blindness.
Read more here: AI for ROP