Phukan, Orchid Chetia, Singh, Ghanapriya, Tiwari, Sanju, and Butt, Saad (2022) An Automated Stress Recognition for Digital Healthcare: Towards E-Governance. In: Communications in Computer and Information Science (1666) pp. 117-125. From: EGERC 2022: First International Conference on Electronic Governance with Emerging Technologies, 12-14 September 2022, Tampico, Mexico.

PDF (Published Version) - Published Version Restricted to Repository staff only

3

Abstract

Mental health is of utmost importance in present times as mental health problems can have a negative impact on an individual. Stress recognition is an important part of the digital healthcare system as stress may act as a catalyst and lead to mental health problems or further amplify them. With the advancement of technology, the presence of smart wearable devices is seen and it can be used to automate stress recognition for digital healthcare. These smart wearable devices have physiological sensors embedded into them. The data collected from these physiological sensors have paved an efficient way for stress recognition in the user. Most of the previous work related to stress recognition was done using classical machine learning approaches. One of the major drawbacks related to these approaches is that they require manually extracting important features that will be helpful in stress recognition. Extracting these features requires human domain expertise. Another drawback of previous works was that it only caters to specific groups of individuals such as stress among youths, stress due to the workplace, etc. and fails to generalize. To overcome the issues related to previous works done, this study proposes a transformer-based deep learning approach for automating the feature extraction phase and classifying a user’s state into three classes baseline, stress, and amusement.

Actions (Repository Staff Only)

Item Control Page