Facial expression detection using posterior-auricular muscle surface-electromyographic activity

Abstract

Introduction: Following Bell's palsy, 25% of sufferers are left with chronic, difficult to treat involuntary muscle movements (synkinesis). Surface-electromyography (sEMG) biofeedback has an established role in improving muscular coordination. The vestigial posterior auricular muscle (PAM) is the first muscle re-innervated by the facial nerve. This study measured the sEMG signal from PAM in facial animation in healthy subjects and explored the possibility of using PAM in a wearable rehabilitative device. Material and methods: PAMs activity was recorded using sEMG in 41 healthy volunteers during the performance of a set range of facial expressions. To compare results between participants and expressions, the percentage increase from baseline were used. Results: PAMs activity increased significantly from baseline in response to smiling (110.98%) compared with other facial expressions: closing eyes (7.85%); blowing out cheeks (39.35%); whistling (11.84%), and showing lower teeth (38.24%). PAMs activity increased in proportion to smile intensity: lips together smile (55.18%); open smile (111.98%); full denture smile (130.07%); and, laughing (269.42%). Conclusion: This study demonstrates that measurement of smile intensity is possible via the sEMG signal from the PAM. Further studies investigating the applicability of measuring PAM activity during smiling to enhance and guide facial palsy rehabilitation are ongoing.

Keywords

• Posterior-auricular muscle
• sEMG
• Surface-electromyography
• Smile
• Facial expression analysis
• Facial palsy rehabilitation