Non-manual grammatical marker recognition based on multi-scale, spatio-temporal analysis of head pose and facial expressions

作者:

Highlights:

• Eyebrow gestures and periodic head movements convey linguistic information.

• We propose multi-scale approach for recognizing non-manual grammatical marker.

• We obtain high-level features from spatio-temporal analysis of non-manual events.

• We use two-stage CRFs to recognize events and partition them into phases.

• Experiments demonstrate superior improvements in ASL.

摘要

•Eyebrow gestures and periodic head movements convey linguistic information.•We propose multi-scale approach for recognizing non-manual grammatical marker.•We obtain high-level features from spatio-temporal analysis of non-manual events.•We use two-stage CRFs to recognize events and partition them into phases.•Experiments demonstrate superior improvements in ASL.

论文关键词:American Sign Language (ASL),Non-manual grammatical markers,Eyebrow height,Head gestures,Facial expressions,Conditional Random Field (CRF)

论文评审过程:Received 7 June 2013, Revised 26 December 2013, Accepted 21 February 2014, Available online 11 March 2014.

论文官网地址:https://doi.org/10.1016/j.imavis.2014.02.009