Computer Vision on a Smartphone: Lessons from Building Clinical-Grade Movement Tracking

Session description

Smartphones are becoming powerful sensing platforms, opening new possibilities for healthcare, accessibility, and human-centered AI. But building reliable computer vision systems outside controlled environments introduces major technical and product challenges.

This talk shares lessons from building a smartphone-based movement tracking platform designed to monitor neurological conditions in children using computer vision and gait analysis.

We’ll explore the engineering challenges of deploying vision models in uncontrolled home environments, dealing with inconsistent lighting and camera angles, ensuring privacy, and balancing accuracy with real-world usability.

Attendees will gain practical insights into deploying edge AI and computer vision in real-world environments where reliability and user trust are critical.