Stress and negative emotions recognition by A@W virtual coach tool

In Ageing@Work, combining measurements from unobtrusive wearables with novel signal processing techniques and advanced machine learning approaches, we detect  stress and negative mood signs detection, in order to allow inference on potential issues that are of major benefit once identified early. Such issues may arise from circadian rhythm disturbances and increased stress levels, possibly related to increased workload and lack of rest, and which may cause harm to the individual either in the short or in the long term.

A specific affective states recognition framework has been developed and integrated in the Ageing@Work virtual coach, that consists of twe s module has been developed and integrated in the Ageing@Work virtual coach tool that consists of two modules: the stress detection module and the emotion recognition module. The former is implemented to shed light into the predictive power of different sensing modalities and classification techniques, while the latter is based on facial expressions in 2D images captured by a camera and is implemented in a smartphone. When the user logs into the A@W App, the camera of the mobile phone is activated and the face detection module is triggered. The detected face is illustrated by the green rectangle. If a good view of the facial image is acquired, the emotion recognition module is executed to identify positive, neutral, or negative emotions visualized in the App with a corresponding smiling, neutral or sad icon face. These instantaneous expressions are aggregated over multiple frames and the most prevailing emotion is stored.

If you want to learn more, check the video below: