An Eye-Tracking Solution Using Consumer Grade Webcams for Potential Concussion Diagnosis and Evaluation

In the past decade, eye-tracking technology has been increasingly utilized in healthcare settings to assess an individual’s cognitive status under various circumstances to evaluate the presence of brain injuries and monitor the evolution of temporary or permanent cognitive illnesses. More specifically, emerging research has confirmed that disconjugate eye movements can be utilized as a predictor of a concussion and potentially return-to-play protocols.Currently, most healthcare applications use eye-tracking devices requiring more detailed signals at a higher sampling rate and with better camera resolution compared to the eye-tracking technology used in the consumer market. Unfortunately, this limits the availability of this type of diagnostic system to clinical settings and prevents its use in situations where early diagnosis is crucial (e.g., during a football game, where 300,000 concussions are reported yearly).In this paper, we introduce a solution for potentially diagnosing and treating concussions based on images acquired with inexpensive and more available camera devices such as webcams, and we detail a performance evaluation study of a popular image segmentation and object detection machine learning model (i.e., MediaPipe Facemesh and Iris) applied to the acquisition and analysis of eye-related signals (e.g., eye movements and blinking) for healthcare applications.