Early AI-Based Diabetic Retinopathy Screening for Bangladesh
Drishti provides accurate Diabetic Retinopathy screening in — all with AI
Drishti is not meant to replace doctors, but rather work in conjuction with specialists to provide initial screening for those who cannot afford full diagnosis. Any diagnosis from Drishti should not be taken as 100% accurate, but should be a guideline for further action. Even if you received a diagnosis of no DR, please be very careful and continue to monitor your eye situation. Continue to get free screenings if possible to ensure you do not progress into developing DR eventually.
Getting Screened by Dristhi is easy — just upload an image and receive a preliminary diagnosis in seconds
We aim to bring AI to Bangladesh where it is rarely utilized. Our AI screening platform is accurate and quick, allowing anyone to get initial screening for free.
Rural areas in Bangladesh are most prone to DR becoming ignored and perpetuated. We deploy our solution specifically in rural areas with affordable hardware to prevent blindness for the most vulernable.
We work with DR specialists, such as ones from the Bangladesh Eye Hospital, as well as multiple organizations to effectively distribute our AI screening to areas where they are needed the most.
We have the support of BRAC University School of Public Health, allowing us to provide reliable and trustworthy verification through clinical research.
Our mission is to provide free, accesible early screening to areas where DR specialists are not available. With academic, clinical, and organizational support, we are able to widen the reach of our service.
Our app does not replace specialists — it is simply just meant to be used in areas where specialists are not readily available and are expensive. Any diagnosis from Drishti should be followed up with a specialist if possible, and if a diagnosis is returned with no DR, please closely monitor your situation and continue to do periodic scannings and diagnoses with Drishti in case your situation gradually becomes worse.
We built a cheap retinal smartphone retinal camera system for cheap and easy retinal imaging. The patent process for this system is in process. In-depth documentation is here.
Moreover, we have published a paper validating the performance of the algorithm on two datasets of Bangladeshi eyes. The paper is available here.