How We Perceive Places: Measuring Biomarkers of Stress in Urban Environments Using Personal Exposure Sensors, Deep Learning and Functional Data Analysis
| Authors | |
|---|---|
| Year of publication | 2024 |
| Type | Article in Proceedings |
| Conference | SAP '24: ACM Symposium on Applied Perception 2024 |
| MU Faculty or unit | |
| Citation | |
| web | https://dl.acm.org/doi/10.1145/3675231.3678876 |
| Doi | http://dx.doi.org/10.1145/3675231.3678876 |
| Keywords | urban environment; stress; high-level visual features; image segmentation; LSTM neural networks; functional data analysis |
| Description | This study investigates the impact of visual stimuli and other environmental factors in urban settings on physiological markers of stress. Forty-four participants walked a predefined urban route while wearable sensors recorded street-level noise, air pollutants, and video footage. Stress responses were measured via heart rate, heart rate variability, and electrodermal activity. Urban scenes were categorized using the SegFormer model, and an LSTM model was applied to predict stress markers. Functional data analysis provided an interpretable model of heart rate responses to environmental variables. Data collection and analyses are ongoing. |
| Related projects: |