Four new papers accepted at ACII 2017 on topics ranging from communicative behavioral patterns to prediction of facial action unit intensities.

July 14, 2017

Members of the MultiComp Lab had four papers accepted at the International Conference on Affective Computing and Intelligent Interaction (ACII 2017, in San Antonio, Texas, USA.

“Local-Global Ranking for Facial Expression Intensity Estimation” by Tadas Baltrušaitis, Liandong Li, and Louis-Philippe Morency introduces the Local-Global Ranking method as a novel alternative to direct prediction of facial action unit intensities; instead of independent frame estimations, the proposed method performs a two-stage ranking involving relative ratings between frames.

“Visual Attention in Schizophrenia: Eye Contact and Gaze Aversion during Clinical Interactions” by Alexandria K. Vail, Tadas Baltrušaitis, Luciana Pennant*, Elizabeth Liebson*, Justin Baker*, and Louis-Philippe Morency (* indicates collaborators at Harvard Medical School) investigates multimodal communicative behavior patterns, particularly those relating to gaze aversion, as indicators of symptoms of schizophrenic symptom severity.

“What Really Matters – An Information Gain Analysis of Questions and Reactions in Automated PTSD Screenings” by Torsten Wörtwein and Stefan Scherer (USC) details an information gain driven analysis of behavioral reactions in automated PTSD screening interviews.

“Hand2Face: Automatic Synthesis and Recognition of Hand Over Face Occlusions” by Behnaz Nojavanasghari*, Charles Hughes*, Tadas Baltrušaitis, and Louis-Philippe Morency (* indicates collaborators at UCF) introduces a novel framework for synthesizing and recognizing naturalistic facial occlusions toward use in the development of more robust facial expression detection models.