========================================================================

BioSensor-Grouplevel-Emotion-Recognition

========================================================================

image

========================================================================

| author | mizukiyuta |

| department | Tokyo Metropolitan University System Design |

========================================================================

Introduction

Use bioSensors for group-level emotion recognition

Used device

[Beacon]

for detecting people in the same room

[Muse Mind Monitor]

Accelerometer ,Gyrometer brain wave(EEG)->Attention,Mediation

[Polar H10]

heart rate(LF/HF)

[Sensor]

BME280 sensor based on raspberry pi4,to get environment data(temperature,humidity,pressure)
image

[VR device]

for emotion Arousal
oculus quest2

Code

[polar H10]use python‘s bleak library to detect polar h10,get its Mac address,and record ECG or Accelerometer.

[PI_SENSOR]BME280 sensor based on raspberry pi4,to get environment data(temperature,humidity,pressure).

[ECG_EEG_PROCESS]EEG and ECG data process program. ECG->lfhf,pnn50... EEG->Attention,Meditation...
process to 60s data,one second one row

========================================================================

Data Collection

Based on VR virtual world to arose people's emotion
meanwhile,sensors will collect collaborators' bio-data and save to csv

Data Processing

Pre-process data,delete no-use parameter and create new parameter
for example:
ECG->LFHF,pnn59... EEG->attetion,meditation...

Group-level utterance atmosphere

1.talk on a certain topic
2.collaborate with each other in VR

Futrure

1.use gaze in VR to recognize interest and attention of dialogue with avater/players
2.consider generlizing model for individual differences

Machine Learning

SVM KNN CNN-SVM LSTM Transformer ...

========================================================================