Representative Image Photo | Express Illustrations
Karnataka

Engineering students in Bengaluru develop AI tool to screen children for eye diseases

According to the team, healthcare workers at these centres could capture images and upload them to the application, allowing the software to flag cases that may require further specialist assessment.

Gayathri M Kurup

BENGALURU: An AI-based software that uses deep-learning technology to analyse eye images and flag possible eye diseases in children was presented by a team of engineering students at the Bengaluru Tech Summit 2026.

The project aims to explore how technology can make preliminary eye screening more accessible, particularly for families in rural areas who face difficulties accessing specialist eye-care services.

Developed as a final-year project by Information Science and and Engineering students Abrar Javaid, Hriday Gautam and Danyal Ibrahim from

Dayananda Sagar Academy of Technology and management (DSATM, the prototype, titled Pediatric Eye Disease Detection Using Deep Learning, is designed to identify 26 eye conditions through image analysis.

The software works by allowing a user to select a patient record and upload a specialised eye image through a web-based application. The image is processed using an EfficientNetB0 deep-learning model, built with TensorFlow and Keras, which generates a prediction across the disease categories configured in the system.

The application also uses Grad-CAM technology to generate a heatmap showing the areas of the image that influenced the model’s prediction. The findings are then organised into a report that can be reviewed by an eye-care professional.

The students said the idea was to address difficulties faced by families in rural and underserved areas, where accessing an ophthalmologist may involve travelling long distances and spending money on consultations and investigations. They envisage the software being made available through local clinics or government-supported health centres equipped with specialised cameras to capture eye images.

According to the team, healthcare workers at these centres could capture images and upload them to the application, allowing the software to flag cases that may require further specialist assessment. This could help avoid unnecessary referrals of children whose images do not indicate a need for immediate specialist review, while helping identify those who need closer examination. The images and reports could also be shared with specialists for remote review, subject to connectivity and appropriate clinical arrangements.

The students said they were exploring hospital associations and possible collaboration with government agencies to take the project forward. They are also working towards clinical validation of the prototype.

‘All of our relationships will improve’: Pakistan minister signals hope for better ties with India

Nobody discriminated against on religious grounds in India, says Rijiju amid Mohammed Siraj row

PLI’s gains, gaps and unfinished agenda

Eight die in three days after consuming suspected spurious liquor in Manipur; advisory issued

SC permits one-time Karthigai Deepam lighting at disputed Thiruparankundram hilltop on Nov 24