Home > Published Issues > 2016 > Volume 4, No. 2, December 2016 >

Stress Recognition from Heterogeneous Data

Bo Zhang, Yann Morère, Loïc Sieler, Cécile Langlet, Benoît Bolmont, and Guy Bourhis
University of Lorraine, Metz, France

Abstract—The assessment of the stress of an individual attracts the attentions of the researchers since it helps to provide individualized assistance in managing this emotional state. This paper investigates the potential of stress recognition using heterogeneous data, where not only the physiological signals but also the Reaction Time (RT) is used to recognize different stress levels. To acquire the data related to mental stress of an individual, we design the experiments with two different stressors: 'Stroop test' and acoustic induction. We develop the classifier based on the Support Vector Machines (SVM) for the stress recognition given the physiological signals. Three physiological signals, Electrodermal Activity (EDA), Electrocardiography (ECG) and Electromyography (EMG), are registered and analyzed. An overall high recognition accuracy of the SVM classifier is obtained. During the experiments, RT task appears. RTs are registered and their statistical analysis shows a generally good discrimination between the period of low stress and the period of high stress. Results indicate that the data from heterogeneous sources, such as physiological signal and cognitive reaction can be adopted for stress recognition.

Index Terms—stress recognition, heterogeneous data, physiological signal, reaction time, Stroop test, acoustic induction

Cite: Bo Zhang, Yann Morère, Loïc Sieler, Cécile Langlet, Benoît Bolmont, and Guy Bourhis, "Stress Recognition from Heterogeneous Data," Journal of Image and Graphics, Vol. 4, No. 2, pp. 116-121, December 2016. doi: 10.18178/joig.4.2.116-121