Our Technology
Our state-of-the-art deep learning algorithm is trained on millions of labeled images to rapidly and accurately assess a variety of health conditions. The camera takes a high resolution image of the fundus — the eyes including the retina and surrounding tissue — getting a high-quality snapshot of both the circulatory and nervous systems. By studying the unique variations among different people, the system is able to distinguish extremely subtle differences that can be translated into an easy-to-understand health assessment.
Benefits
![](https://static.wixstatic.com/media/18f3b7_98bf6b9ad7eb405bb178eb4563cf6035~mv2.jpeg/v1/fill/w_475,h_299,al_c,q_80,usm_0.66_1.00_0.01,enc_auto/18f3b7_98bf6b9ad7eb405bb178eb4563cf6035~mv2.jpeg)
Our Science
Dozens of high-quality academic papers (here's a list) validate the underlying technology of using retinal images.
​​
The key research paper was published in Aug 2021 in The Lancet and showed an AUC of 0.955 when evaluated on 200K fundus photos. The study was sponsored by Sun Yat-sen University and pre-registered at Clinicaltrials.gov.
Benefits
For Patients
-
Painless: a brief, low-intensity flash in each eye. No poking, no messy fluids.
-
Fast: comprehensive results available in under five minutes
-
Easy: follow the voice prompts for automatic scanning.
​
For Practitioners
-
Fully-automatic: self-guided voice prompts let anyone use it unassisted with no training required
-
Portable: fits in a carryon bag and can run on batteries. Ideal for mass screening away from the clinic.
-
Flexible reporting: onboard patients using their smartphone for self-operation. Or use the optional clinical console to enter patient information and receive reports directly for later discussion with patients.
-
Low-cost: an order of magnitude lower. A similar-resolution medical grade fundus camera typically costs over $50,000 whereas ours is available for under $10K.
​
![](https://static.wixstatic.com/media/18f3b7_e82d2e7315f245678362996073b339d2~mv2.webp/v1/fill/w_482,h_366,al_c,q_80,usm_0.66_1.00_0.01,enc_auto/18f3b7_e82d2e7315f245678362996073b339d2~mv2.webp)