Emotion Recognition using Wireless Signals

Emotion Recognition using Wireless Signals

Mingmin Zhao, Fadel Adib, Dina Katabi. International Conference on Mobile Computing and Networking (Mobicom’16)

Abstract

This paper demonstrates a new technology that can infer a person’s emotions from RF signals reflected off his body. EQ-Radio transmits an RF signal and analyzes its reflections off a person’s body to recognize his emotional state (happy, sad, etc.). The key enabler underlying EQ-Radio is a new algorithm for extracting the individual heartbeats from the wireless signal at an accuracy comparable to on-body ECG monitors. The resulting beats are then used to compute emotion-dependent features which feed a machine-learning emotion classifier. We describe the design and implementation of EQ-Radio, and demonstrate through a user study that its emotion recognition accuracy is on par with stateof-the-art emotion recognition systems that require a person to be hooked to an ECG monitor.