Emotion Recognition Using Footstep-Induced Floor Vibrations

Researchers in the Noh Lab have developed a gait based, emotion recognition system using geophone sensors that are attached to the floor. People's gait changes under various emotions creating distinct structural vibration patterns. The gait-based emotion recognition system collects and analyzes those patterns in three modules: footstep detection and data preprocessing; emotion-related feature extraction; and emotion recognition, which is a multilayer perceptron model to estimate the pedestrians' emotional states. This approach allows for a non-invasive emotion recognition with applications in mental health monitoring, human-computer interaction, emotion-driven advertisement, and to provide personalized and relevant suggestions from recommendation systems.

Stage of Development – Proof of Concept
The Noh Lab continues to test the robustness of the hardware and software under different environments and to develop and refine the user interface.