Oculomotor and facial patterns in video monologues: associations with content characteristics and speech rate
The project is aimed at automated analysis of non-verbal behavior in public video monologues: gaze metrics (vertical/horizontal) and blink rate synchronized with the transcript by timecodes at the level of speech segments. The study examines the relationship of these indicators with the characteristics of the content of utterances (temporal orientation, modality) with the control of speech tempo. The result is a reproducible pipeline of multimodal analysis and empirical estimates of associations between speech parameters and nonverbal markers.
