iCrutch

Acquired nystagmus is a condition characterized by repetitive, involuntary eye movements that impair visual stability and quality of life. Current treatments, including pharmacological therapy and surgery, are often ineffective, non-specific, and unable to adapt to progressive symptom changes. Here, we present a real-time eye-tracking system that detects nystagmus, classifies its subtype using machine learning, and computes a corrective motion vector. The system integrates infrared cameras with Fourier-based processing to distinguish pathological oscillations from voluntary gaze shifts, achieving 99% detection accuracy with a 10 ms response latency. A computational model predicts the necessary counteracting motion to stabilize gaze, mapping corrective movements to extraocular muscles. Preliminary results demonstrate high precision in nystagmus classification and motion compensation, establishing a foundation for future electrode-based stimulation therapies. This system represents a step toward adaptive, closed-loop therapies.

Project video