Machine-based Emotion and Stress Recognition – Towards a “Mood-aware Internet of Things”?

Imagine smartphones or vehicles reacting to user’s moods… social robots showing compassion for a sad frail human… emotions reliably monitored by a remote healthcare app…

The Mood-aware Internet of Things, as the affectiva blogemphasises, can become the cornerstone of a new world in which technology and humans come closer to each other than ever before.

Emotions are often used as synonyms for feelings and effects. According to the field literature, they represent a combination of physiological excitement (such as palpitations, sweat, muscle tension), motoric expression (such as open mouth and eyes, faltering breath, muscle tension, etc.), tendency to act (like the urge to run away) and subjective feeling (e.g. the feeling of being in danger) (Scherer, 2002, p. 166).

The concept of stress commonly designates the reality of being under pressure or strain. According to Selye (the father of stress research), the reaction to stress in humans and animals consists of certain phases: alarm, resistance and exhaustion (Selye cited in Rensing, Koch, Rippe, & Rib, 2005, p.5). It is important to distinguish between eustress (the positive, stimulating aspects of short-term stress) and distress (causing desperation, exhaustion, misery, cases in which the physiological reaction reaches a disease-promoting state)

Machine-based recognition of stress and emotions has become an important topic of technological research and development. Recent advances in sensors and wearable computing have enabled the detection of conditions such as joy, excitement, reverie or lack of concentration on the basis of various sensor data (such as EEG, ECG, breathing, movement, voice, facial expressions recorded by cameras).  Concerning machine-based facial recognition alone, it has been estimated that the global market will grow from $2.77 billion in 2015 to $6.19 billion in 2020. For an overview of the latest emotion recognition apps based on this, see 20+ Emotion Recognition APIs That Will Leave You Impressed, and Concerned,(Bill Doerrfeld | 31 December 2015, updated on 11 August 2016)

The possibility for reliable recognition and prognosis of the emotional states enables, in principle, better decision-making in many usage contexts (driving, work, healthcare, etc.). We are aware of several projects by Silicon Alps members relating to technology and emotions:

We are looking forward to learning more about captivating usage cases and projects. Please don’t hesitate to contact usif you do research and development on sensor-based emotion and stress recognition.



Rensing, L., Koch, M., Rippe, B., Rippe, V. (2013), Mensch im Stress. Psyche, Body, Molecules, Springer Spektrum Publishing

Rana el Kaliouby, Gabi Zijderveld, The Mood-Aware Internet of Things (blog),, AFFECTIVA, 24/07/15

Bill Doerrfeld, 20+ Emotion Recognition APIs That Will Leave You Impressed, and Concerned (Blog), 31 December 2015, updated on 11 August 2016)

Scherer K. (2002) Emotion. In: Stroebe W., Jonas K., Hewstone M. (eds) Sozialpsychologie [Social Psychology]. Springer-Lehrbuch. Springer, Berlin, Heidelberg