Facial paralysis modeling based on image morphing

Abstract

Facial Paralysis is a serious disease that involves the paralysis of any structures innervated by the facial nerve and requires immediate treatment. Yanagihara method is a popular method for the measurement of facial paralysis in Japan. In this method, the patient is asked to make 10 different expressions and the doctor applies the score to each expression. Typical sample expression facial images for different scores are important for educations and references. Since it is not possible to use patient's images as typical sample images, we propose a facial paralysis modeling method based on image warping to generate a virtual expression image of facial paralysis as same as real expression of facial paralysis. We transform the movement or expression of the patient to an average facial image by using both B-spline warping and piece-wise affine transform.