ID 2510: Video-based digital biomarker extraction framework for psychological research

Bachelor’s / Master’s Thesis / Research Internship / Project

Psychological assessments often rely on subjective measures such as self-reports and interviews. However, recent advancements in computer vision and machine learning enable the extraction of objective biomarkers from video data, such as facial expressions, micro-expressions, gaze patterns, speech and verbal acoustics, as well as body posture and movements. For that reason, we are currently building a software framework that unifies all these different modalities and will provide an open-source, modular, scalable and flexible video-based analysis framework for various topic in psychology (stress assessment, affect recognition during sleep, ).

This thesis will focus on contributing to this framework to extract and analyze such biomarkers, helping to advance digital health psychology solutions, and to analyze existing video data from us or from our collaborators.

Tasks

• You will work in an interdisciplinary team with psychologists and engineers
• You have the opportunity to apply your programming and data science knowledge to practical applications in psychology, especially stress research
• You have the unique opportunity to contribute to a growing data analysis framework for video-based digital biomarker analysis. This includes:
  • Implementing different algorithm blocks into an existing framework
  • Learning how to write reusable, modular, and well-structured code
  • Performing statistical and machine learning-based data analysis on data from existing studies

Requirements

• Knowledge in data science and machine learning in Python, including data processing, data analysis, and data visualization
• Interest in open science, affective computing, and psychology, especially biological psychology, psychophysiology, and behavioral analysis

Supervisors

Robert Richer, M. Sc. PhD Candidate & Group Head