Comparison of the DRISTi -Artificial Intelligence System with Ophthalmologists’ Grading for Diabetic Retinopathy Detection

India currently has 101 million people with diabetes, expected to rise to 134 million by 2045. Diabetic retinopathy affects about 35 percent of people with diabetes for over five years. Due to limited eye care access and few ophthalmologists, only a small percentage of diabetics receive regular screening. AI-based tools  can  help scale DR detection. This study aims to compare the diagnostic performance of Artificial Intelligence System  ( called DRISTi) with that of expert ophthalmologists.  The primary objective of the study is  compare the sensitivity and specificity of DRISTi in detecting  Referable  Diabetic Retinopathy  from  fundus images versus ophthalmologist grading.The study is designed as a Retrospective observational study to be done  over  maximum Three months for data collection, analysis, and reporting. An estimated  2500  images   from  different non-mydriatic cameras (Crystal Vue, )    will be de-identified and anonymided .This data set will be secure ly stored and garded by both AI and the  Ophthalmologist . The opthalmologist will be  blinded to AI results. Statistical anlaysis  will include  snsitivity and specificity in detecting referable DR awith  95 percent confidence intervals .