Automated tracking of facial features in patients with facial neuromuscular dysfunction

Facial neuromuscular dysfunction severely impacts adaptive and expressive b ehavior and emotional health. Appropriate treatment is aided by quantitativ e and efficient assessement of facial motion impairment. We validated a new ly developed method of quantifying facial motion, automated face analysis ( AFA), by comparing it with an established manual marking method, the Maxima l Static Response Assay (MSRA). In the AFA, motion of facial features is tr acked automatically by computer vision without the need for placement of ph ysical markers or restrictions of rigid head motion. Nine patients (seven w omen and two men) with a mean age of 39.3 years and various facial nerve di sorders (five with Bell's palsy, three with trauma, and one with tumor rese ction) participated. The patients were videotaped while performing voluntar y facial action tasks (brow raise, eye closure, and smile). For comparison with MSRA, physical markers were placed on facial landmarks. Image sequence s were digitized into 640 x 480 x 24-bit pixel arrays at 30 frames per seco nd (1 pixel congruent to 0.3 mm). As defined for the MSRA, the coordinates of the center of each marker were manually recorded in the initial and fina l digitized frames, which correspond to repose and maximal response. For th e AFA, these points were tracked automatically in the image sequence. Pears on correlation coefficients were used to evaluate consistency of measuremen t between manual (the MSRA) and automated (the AFA) tracking methods, and p aired/tests were used to assess the mean difference between methods for fea ture tracking. Feature measures were highly consistent between methods. Pea rson's r = 0.96 or higher, p < 0.001 for each of the action tasks. The mean differences between the methods were small; the mean error between methods was comparable to the error within the manual method (less than 1 pixel). The AFA demonstrated strong concurrent validity with the MSRA for pixel-wis e displacement. Tracking was fully automated and provided motion vectors, w hich may be useful in guiding surgical and rehabilitative approaches to res toring facial function in patients with facial neuromuscular disorders.