Written by Third Place Team EyeConic
When we joined the opening session of the Orbis Future Vision Leaders Visionary Innovators Hackathon, most of us were meeting for the first time. Our team included students and trainees from Canada, Hong Kong, and Ghana—medical students, nursing students, ophthalmology residents, and optometry students. Despite our different backgrounds and time zones, we were all drawn to the event for the same reason: Orbis’s mission to make quality eye care accessible to everyone.
The hackathon began with discussions on global ophthalmology and the gaps that still limit access to early diagnosis and treatment. We were then assigned our challenge prompt: amblyopia, a condition that remains one of the leading causes of preventable vision loss in children. Early signs can be subtle and easily missed, especially in settings without consistent access to pediatric eye specialists. That prompt immediately gave our team a clear and meaningful focus.
Early risk factors for amblyopia can be surprisingly subtle—slight eye misalignment, drooping of one eyelid, or unequal light reflexes may not cause obvious symptoms. Young children often can’t articulate that something is wrong, and in many communities, routine vision checks aren’t readily available. As a result, early warning signs are frequently overlooked until vision loss has already developed.
As the week went on, our team quickly found common ground in how we wanted to approach the problem. We decided to design a tool that could help bridge the early-screening gap in a practical, scalable way. This became the foundation of PediaSight.
PediaSight is a smartphone-based AI tool we built to help detect early risk factors for amblyopia including strabismus, red-reflex asymmetry, ptosis, and refractive indicators, using a convolutional neural network trained on pediatric ophthalmic images. We paired the model with a simple, guided image-capture process and clear risk outputs so the tool could be used in schools, primary care settings, community programs, or even at home. Our aim was to create something that could realistically support earlier referrals for children who might otherwise be missed.
PediaSight helps close this gap by guiding users to take a simple photograph of a child’s eyes and instantly analyzing it for early indicators of concern. Instead of relying on specialized equipment or expert interpretation, the app flags patterns that may suggest risk and provides clear recommendations on when to seek professional care. In other words, it turns a smartphone into an accessible first-step screening tool that can support earlier detection anywhere.
By the end of the hackathon, we had built a functioning model, a user-friendly workflow, and a vision for how PediaSight could support early detection in diverse settings. The experience showed us how much can be accomplished when multidisciplinary perspectives come together with a shared goal, and it strengthened our commitment to continuing this work beyond the event.
As a team, we hope to refine PediaSight further and explore opportunities to pilot it in practice. We’re motivated by the potential to expand access to early pediatric eye screening—one step toward the broader vision of equitable eye care that Orbis champions around the world.
EyeConic Team:
Dr. Amaar Amir, Canada
Ms. Angelica Hanna, Canada
Ms. Kyobin Hwang, Canada
Ms. Tasha Miller, Canada
Mr. Andrew Mihalache, Canada
Mr. Tsz Yun Chan (Peter), Hong Kong
Ms. Constance Appiah Boahemaa, Ghana
Ms. Ka Hei Leung (Katherine), Hong Kong
