Tobias Gessler

Advisors

Richer Robert (M. Sc.), Luca Abel (M. Sc.), Arne Küderle (M. Sc.), Prof. Dr. Björn Eskofier

Duration

03 / 2023 – 08 / 2023

Abstract

Stress is a complex phenomenon that affects individuals in various ways, impacting both psychological and physiological states [1, 2]. Previous research has shown that exposure to acute psychosocial stress can significantly influence body posture and movement [2], providing a promising extension to the existing markers for acute stress, such as cortisol and inflammatory markers [3]. Traditionally, motion analysis is performed using optical or inertial measurement unit (IMU)-based motion capturing. Although these techniques offer the highest accuracy, they come with high costs, limited availability, and can interfere with natural behaviour during stress induction. A more feasible alternative is video-based motion analysis using tools like OpenPose [4] or AlphaPose [5]. They provide a cost-effective and straightforward way to extract and analyze motion patterns based solely on video data. However, despite these advantages, video-based motion extraction and analysis have not yet been widely adopted for psychosocial stress analysis. This is likely due to the limited accessibility of those tools for non-technical researchers. In comparison to facial expression analysis where tools such as FaceReader [6] provide a convenient solution, there is a lack of out-of-the-box solutions for motion analysis.
The goal of this bachelor’s thesis is therefore to develop a web framework for video-based motion analysis, with a particular focus on acute psychosocial stress. The platform shall be capable of performing large-scale end-to-end extraction of motion parameters. One aim of the web platform is to provide non-technical researchers with an easy-to-use tool to extract meaningful motion information from videos acquired during acute psychosocial stress. The platform will be developed with a modular architecture, enabling expansion and adaptation of the processing pipelines to meet evolving research requirements. Additionally, to demonstrate the capabilities of the OpenTSST framework, a proof-of-concept analysis will be performed on data from an existing study that investigated the effect of stress habituation to repeated acute psychosocial stress. The existing analysis will be extended by investigating whether habituation effects can not only be observed in the cortisol response to repeated acute stress, but also in the freezing-related motion parameters.

 

Full Thesis

 

References

[1] D. B. O’Connor, J. F. Thayer, and K. Vedhara, “Stress and Health: A Review of Psychobiological Processes,” Annu. Rev. Psychol., vol. 72, no. 1, pp. 663–688, 2021, doi: 10.1146/annurev-psych-062520-122331

[2] L. Abel, “Machine Learning-Based Detection of Acute Psychosocial Stress from Dynamic Movements,” Master’s Thesis in Medical Engineering, Friedrich-Alexander-Universität Erlangen-Nürnberg

[3] Y. M. Ulrich-Lai and J. P. Herman, “Neural regulation of endocrine and autonomic stress responses,” Nat. Rev. Neurosci., vol. 10, no. 6, pp. 397–409, Jun. 2009, doi: 10.1038/nrn2647

[4] Z. Cao, G. Hidalgo Martinez, T. Simon, S. Wei, and Y. A. Sheikh, “OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019.

[5] H. Fang, J. Li, H. Tang, C. Xu, H. Zhu, Y. Xiu, Y. Li, and C. Lu, “AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

[6] P. Lewinski, T. M. Den Uyl, and C. Butler, “Automated facial coding: Validation of basic emotions and FACS AUs in FaceReader,” Journal of Neuroscience, Psychology, and Economics, vol. 7, no. 4, pp. 227, 2014, Educational Publishing Foundation