Date: November 2015.
Source: 2015 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Pages 337-342. Washington, DC.
Abstract: Autism Spectrum Disorder (ASD) impairs an individual’s non-verbal skills including natural and contextual facial expressions. Such impairments may manifest as odd facial expressions (facial oddity) based on subjective evaluations of facial images. A few studies conducted on individuals with ASD have focused on the physiology of facial muscle usage by employing eletrophysiological sensors in response to visual stimuli. The sensors are placed directly on the face and may inhibit or limit the spontaneous facial response which may be too subtle for subjective human evaluations. This study uses a non-intrusive 3D facial imaging sensor that captures detailed geometric information of the face to facilitate quantification and detection of subtle changes in facial expression based on the physiology of facial muscle. A novel computer vision and data mining approach is developed from curve-based geometric feature of 3D facial data to discern the changes in the facial muscle actions. A pilot study is conducted with sixteen subjects (8 subjects with ASD and 8 typically-developing controls) where 3D facial images have been captured in response to visual stimuli involving 3D facial expressions. Statistical analyses reveal a significantly asymmetric facial muscle action in subjects with ASD compared to the typically-developing controls. This study demonstrates feasibility of using non-intrusive facial imaging sensor data in evaluating possible physiology-based impairments.
Article: Analysis of facial muscle activation in children with autism using 3D imaging.
Authors: Samad, Manar D; Bobzien, Jonna L; Harrington, John W; Iftekharuddin, Khan M. Department of Electrical and Computer Engineering, Old Dominion University, Norfolk, VA USA.