A prototype for automatic image analysis to quantify rehabilitation of chronic facial paralysis

Resumen

A prototype computerized system was developed to recognize facial features like the pupils of the eyes and the corners of the mouth in video images of patients with facial paralysis. The system uses image processing and the Hough transform for facial feature identification. Once these features are obtained, it can measure the distance between them or to a reference point in the image. The method has been tested with examples of real images of persons undergoing therapy with good results. The purpose of the system is to assist in an objective, quantitative rehabilitation in which the patients regain symmetry of their facial musculature and regain control over their facial expressions. This work is important because quantified and objective measures of rehabilitation progress are not widely used in therapy. Semi-automated computer techniques promise to not only provide more accurate and effective therapy but also to lower the cost of therapy and may lead to systems that are affordable for home use.