Emotion Recognition Lab
The work in this lab aims at recognizing human activities by fusing various sensor types.
Emotions have implications on our past, present and future living. They modify what information we do remember, how we think and decide and how we plan our next steps. Understanding and interpreting emotions in a given context is therefore a powerful tool to have access to underlying intentions and motivations. Since strong emotions change physical as well as behavioral parameters, a combination of different IoT sensors can be used to identifying and quantifying human emotion. We implement our emotion recognition methods by fusing different cues including audio (human voice), physiological parameters (heartbeat, sweat level, breathing rate) and motion sensors. Sophisticated machine learning methods are required to derive e.g. arousal (emotion intensity) and valence (positive or negative emotion) from the extracted features.
From our lab team member:
“Applying signal processing and machine learning algorithms to detect emotional responses in various situations has always been fascinating me. Emotion Recognition is bridging the gap between technology and humanity and we are just at the beginning of these techniques materializing and bringing AI to a new level. IoTA is the place where I’m pushing as a Data Scientist emotion recognition methods from the lab to real working solutions.”