TRACE: Transparent Recognition of Affective Cues via Expression
Faculty Mentor
Dr. Felix Hamza-Lup
Location
Ogeechee Theater
Type of Research
On-going
Session Format
Poster Presentation
College
Allen E. Paulson College of Engineering & Computing
Department
Computer Science
Abstract
Microexpressions are brief, involuntary facial expressions lasting 40–200 milliseconds that occur when individuals attempt to conceal emotions. Unlike conventional facial expressions, which typically last between 0.5-4 seconds, microexpressions emerge automatically and reveal authentic affective states largely resistant to conscious control. This involuntary quality makes microexpressions a powerful signal for research in clinical diagnosis, security screening, behavioral analysis, and human–computer interaction applications. From a computing perspective, microexpression analysis presents an exceptional challenge: signals are subtle, temporally sparse, and highly individualized, while labeled training data remain scarce. These constraints demand advances in spatiotemporal modeling, fine-grained feature extraction, and data-efficient learning methods. Our research focuses on precise mapping of Facial Action Units to discrete emotional states, enabling models to capture causally meaningful facial dynamics rather than surface-level correlations. This research advances computing toward human-aware, transparent, and socially beneficial intelligent systems grounded in measurable, biologically meaningful facial behavior, with implications for science and society.
Program Description
.
Start Date
4-21-2026 11:00 AM
End Date
4-21-2026 11:15 AM
Recommended Citation
Huynh, Thao Ngan, "TRACE: Transparent Recognition of Affective Cues via Expression" (2026). GS4 Student Scholars Symposium. 47.
https://digitalcommons.georgiasouthern.edu/research_symposium/2026A/2026A/47
TRACE: Transparent Recognition of Affective Cues via Expression
Ogeechee Theater
Microexpressions are brief, involuntary facial expressions lasting 40–200 milliseconds that occur when individuals attempt to conceal emotions. Unlike conventional facial expressions, which typically last between 0.5-4 seconds, microexpressions emerge automatically and reveal authentic affective states largely resistant to conscious control. This involuntary quality makes microexpressions a powerful signal for research in clinical diagnosis, security screening, behavioral analysis, and human–computer interaction applications. From a computing perspective, microexpression analysis presents an exceptional challenge: signals are subtle, temporally sparse, and highly individualized, while labeled training data remain scarce. These constraints demand advances in spatiotemporal modeling, fine-grained feature extraction, and data-efficient learning methods. Our research focuses on precise mapping of Facial Action Units to discrete emotional states, enabling models to capture causally meaningful facial dynamics rather than surface-level correlations. This research advances computing toward human-aware, transparent, and socially beneficial intelligent systems grounded in measurable, biologically meaningful facial behavior, with implications for science and society.
