Finding mannerisms without labels
A sparsity-based algorithm extracts recurring movement patterns that independent observers judged far more accurate than random selection. The interface pairs each pattern with a skeleton animation and simple questions.
-
AutoManner: An Automated Interface for Making Public Speakers Aware of Their Mannerisms
Shows public speakers their own repetitive mannerisms, automatically found in their videos.
-
Unsupervised Extraction of Human-Interpretable Nonverbal Behavioral Cues in a Public Speaking Scenario
Finds interpretable nonverbal cues in public-speaking videos without labels.
