[4G2-OS-24c] 日常生活知識とAI

[4G2-OS-24c-02] Stress Analysis in Remote Work Using Multimodal Sensing

Daiki Kishimoto¹, Takumi Sugiura¹, Yuji Yamada¹, 〇Kazuki Ito², Hiroki Aramaki², Satoshi Kurihara³ (1. Graduate School of Science and Technology, Keio University, 2. Net One Systems Co., Ltd., 3. Faculty of Science and Technology, Keio University)

Keywords:Multimodal Sensing, Human Sensing, Data Science

In a remote work environment, where communication is mainly online, it is difficult to understand stress through communication. Therefore, it is important to detect stress among remote workers while they continue their work. In this study, we focused on PC operation log analysis and facial expression analysis as approaches to this problem. The purpose of this study is to clarify the relationship between stress of remote workers and PC operation features and the estimated emotional state based on facial expression analysis. First, an environment was constructed to acquire data from a heart rate sensor, a PC operation log acquisition program, and a web camera, and to process and integrate the data. After setting the tasks, we conducted simulated office work and remote work to acquire the data. As a result, it suggested that the number of left clicks and the emotional estimates of disgust and sadness were related to changes in stress, but all the correlations were weak. However, the results suggest that stress can be reduced by using office work instead of remote work until the employees become accustomed to the work, a finding that may be important in implementing remote work.