Archive · 30 Jan 2026
AI Eye Scan Detects Diabetes Without Blood Tests
Scientists from India and the US have created a smart computer program that can find diabetes by taking a picture of the back of your eye.
This is cool because it doesn't need a blood test.
The computer looks for tiny signs in your eye's blood vessels that people can't see.
They tested it on pictures from 139 people and it was really good at finding diabetes.
This could help lots of people in India, where many have diabetes but don't know it.
The best part is that it's quick and doesn't hurt.
The scientists want to test it on more people to make sure it works well.
Researchers developed an AI technique to detect diabetes using retinal photos.
The method analyzes tiny signs in eye blood vessels invisible to the human eye.
Study published in Diabetes Technology and Therapeutics showed 95% accuracy in detecting diabetes.
The AI was trained to identify specific shapes and patterns in veins using retinal images.
The non-invasive technique could help in early diabetes detection without blood tests.
- Who
- Indian and US researchers, including Dr. V. Mohan and Dr. Sudeshna Sil Kar
- What
- Developed an AI-based technique to detect diabetes using retinal photos
- When
- Study published in the Diabetes Technology and Therapeutics journal
- Where
- Research conducted by teams from India and the US, including Yenepoya University in Karnataka
- Why
- To provide a non-invasive, early detection method for diabetes without the need for blood tests
- Research Team
- Indian and US researchers
- Detection Method
- High-resolution retinal photo
- Study Journal
- Diabetes Technology and Therapeutics
- Sample Size
- 273 retinal images from 139 participants
- Accuracy
- 95% sensitivity in detecting diabetes
- Key Contributor
- Dr. V. Mohan, Chennai-based diabetologist
- AI Training Focus
- Specific shapes and patterns in veins