OpenFaceR: Developing an R Package for the Convenient Analysis of OpenFace Facial Information.

EasyChair Preprint 4389

Abstract

OpenFace is an open source tool designed to extract the most commonly used facial information from videos including facial points, head pose, gaze and Facial Action Units. OpenFaceR is a tool designed to help social scientists, and other researchers from less technical disciplines, who are interested in facial nonverbal behaviors (FNVBs), to easily use output from OpenFace 2.0. The output from OpenFace is one csv file for each video, with information on each feature for each frame of the analyzed video provided in rows.  OpenFaceR constitutes a set of methods to convert information in this format into relevant summary statistics. In this paper, we focus on the set of methods in OpenFaceR to extract information from a series of videos and transform the output files into a single dataset in which each row reports the summary values of a feature for one video.

Keyphrases: OpenFace, R package, computer vision, face, nonverbal behaviors

@booklet{EasyChair:4389,
  author    = {Davide Cannata and Sam Redfern and Denis O'Hora},
  title     = {OpenFaceR: Developing an R Package for the Convenient Analysis of OpenFace Facial Information.},
  howpublished = {EasyChair Preprint 4389},
  year      = {EasyChair, 2020}}