Facial Plastic and Reconstructive Surgery

Facial palsy describes the clinical problem of unilateral facial paralysis or weakness. It is surprisingly common, with approximately 150,000 cases per year in the United States. Despite this, clinical scoring of paralysis severity is cumbersome. Given recent advances in AI-driven computer vision, it is feasible to employ computer vision for facial landmark detection and track movement of these landmarks over time. These data can then be used to score the severity of facial paralysis. If accurate in comparison to human landmark annotations, these measures will be compared to trained clinician scores of disease severity. In its final form, this algorithm will permit remote, accurate analysis of facial paralysis. This is an important step forward in clinical triage and gauging the efficacy of different treatments for this disease.