Inspiration
People with visual impairments often navigate environments without access to the visual information others take for granted. We wanted to turn technology into an intelligent “extra set of eyes” that helps users understand their surroundings and move with greater confidence and independence.
What it does
Sightra uses computer vision and AI to detect and interpret objects, obstacles, and potential hazards in real time. It converts this visual information into contextual audio guidance delivered through connected earbuds, helping users make safer navigation decisions.
How we built it
Sightra combines a camera-based sensing system, computer vision for real-time environmental detection, AI for contextual interpretation, and IoT connectivity for the audio interface. The system processes the user's surroundings and delivers concise spoken guidance through Bluetooth-connected earbuds.
Challenges we ran into
The biggest challenges were achieving low-latency vision processing, generating useful contextual descriptions rather than simple object labels, and designing the system to remain practical on constrained hardware while providing clear audio feedback without overwhelming the user.
Accomplishments that we're proud of
We built a working prototype that can perceive and interpret real-world surroundings and communicate actionable information through audio. We are particularly proud of combining AI, computer vision, and connected hardware into an assistive system focused on a tangible accessibility problem.
What we learned
We learned that assistive AI must prioritize clarity, latency, reliability, and user context over simply detecting more objects. The most useful output is not everything the system sees, but the information that helps the user make a decision.
What's next for Sightra
Next, we plan to move toward a more compact wearable device, improve real-time and edge processing, add richer spatial awareness and hazard detection, expand support for African environments and languages, and conduct structured testing with visually impaired users to refine Sightra around real-world needs.
Log in or sign up for Devpost to join the conversation.