i_mBODY Lab

Interactive Multisensory Body-centred Experiences
at the Intersection of Neuroscience & Technology


Understanding human affect through non-verbal cues

Date: 16th November 2022,

Speaker: Prof. David Masip, Universitat Oberta de Catalunya, AI-WELL laboratory.


Humans communicate our emotions using non-verbal language. One of the most studied affective computing informational cues is the analysis of facial expressions. Nevertheless, faces convey far more information than the 6 classical emotions defined in psychology. In this talk, I will introduce two research ideas developed in the Ai-Well group:

  • Attitudes prediction from (very) subtle spontaneous facial expressions in short videos, where we try to predict the binary preferences from covert videos of the observer’s faces.
  • Perception of personality traits from videos, where we explore the feasibility of predicting the “Big five” personality traits (openness, conscientiousness, extraversion, agreeableness, and neuroticism) from short video interactions.


David Masip is professor in the Computer Science Department at Universitat Oberta de Catalunya since February 2007, and director of the Doctoral School since 2015. He studied Computer Science at the Universitat Autonoma de Barcelona, obtaining an FPI grant in 2001 to start his Ph.D. degree in the Computer Vision Center (Spain). He obtained the Ph.D. degree in September 2005. He obtained the best thesis award in computer science. Previously, he worked as an assistant professor in the Applied Mathematics Department at the Universitat de Barcelona, and in 2013 he was a Research Affiliate at Princeton University. He advised 3 doctoral theses (and 4 ongoing) and participated in 14 research projects, being the PI on 4 of them. He published 28 high-impact journal publications and more than 30 conference papers.

His research interests are in computer vision and machine learning, particularly Deep Learning (DL) algorithms applied to Affective Computing and Medical Image Analysis in eHealth. His current projects focus on adding explainable layers to DL, training DL models in small sample size problems (transfer and self-supervised learning), uncertainty modeling in DL, and applications of DL models to Retina Imaging.

 If you are interested in giving a talk please write an email to: lab.imbody@gmail.com.