An Unsupervised Approach for Study of the Relationship Between Behavioral Profile and Facial Expressions in Video Interviews

Ana Conceição de Jesus, Richard Mariano, Alessandro Vieira, Jéssica Almeida de Lima, Giulia Zanon de Castro, Wladmir Brandão, Wladmir Brandão

2023

Abstract

The use of Behavior Mapping (BM) questionnaires to analyze the Behavioral Profile (BP) of employees can lead to biased responses, incompleteness, and inaccuracies, especially when information such as reactions, facial expressions, and body language cannot be captured. Thus, methods with the ability to recognize and characterize BP automatically and without direct inference of bias can minimize the impact of erroneous assessments. Intensity clustering of Facial Action Units (AUs) from the Facial Action Coding System (FACS), extracted from video interviews of PACE BPs, was proposed. Features were extracted from 500 videos and effort was targeted to profiles whose probability was equal to or greater than 40% of the individual belonging to the profile. An analysis of the relationship between BPs and intensities throughout the video was presented, which can be used to support the expert’s decision in the BM.

Paper Citation

in Harvard Style

Conceição de Jesus A., Mariano R., Vieira A., Almeida de Lima J., Zanon de Castro G. and Brandão W. (2023). An Unsupervised Approach for Study of the Relationship Between Behavioral Profile and Facial Expressions in Video Interviews. In Proceedings of the 25th International Conference on Enterprise Information Systems - Volume 1: ICEIS, ISBN 978-989-758-648-4, SciTePress, pages 493-500. DOI: 10.5220/0011856100003467

in Bibtex Style

@conference{iceis23,
author={Ana Conceição de Jesus and Richard Mariano and Alessandro Vieira and Jéssica Almeida de Lima and Giulia Zanon de Castro and Wladmir Brandão},
title={An Unsupervised Approach for Study of the Relationship Between Behavioral Profile and Facial Expressions in Video Interviews},
booktitle={Proceedings of the 25th International Conference on Enterprise Information Systems - Volume 1: ICEIS,},
year={2023},
pages={493-500},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011856100003467},
isbn={978-989-758-648-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 25th International Conference on Enterprise Information Systems - Volume 1: ICEIS,
TI - An Unsupervised Approach for Study of the Relationship Between Behavioral Profile and Facial Expressions in Video Interviews
SN - 978-989-758-648-4
AU - Conceição de Jesus A.
AU - Mariano R.
AU - Vieira A.
AU - Almeida de Lima J.
AU - Zanon de Castro G.
AU - Brandão W.
PY - 2023
SP - 493
EP - 500
DO - 10.5220/0011856100003467
PB - SciTePress