EmotiphAI_public

EmotiphAI is a platform developed to address the challenge of collecting physiological data from groups, particularly when a centralised controller is used.

Motivation

The platform is designed not only for real-time biosignal acquisition but also for retrospective emotion annotation. By analyzing Electrodermal Activity (EDA) data, EmotiphAI identifies significant moments in a session (e.g., during a 2-hour movie), allowing for targeted annotation. This approach minimizes distraction during the emotion elicitation process, making it more efficient and user-friendly.

Methods

EmotiphAI is built on a low-cost, standalone local infrastructure, which includes:

• Hardware:

  • A local hub, such as a Raspberry Pi or Odroid, that serves as the central data receiver.
  • A wearable device, 3D-printed and based on the ESP32 microcontroller, for biosignal acquisition.

• Communication:

  • Data is transmitted via Bluetooth to the local hub, which is connected through a WiFi router (e.g., TP-Link Wireless N 450Mbps (TL-WR940N)).
  • Multiprocessing is employed to manage simultaneous data reception from multiple devices while optimizing CPU core usage.

• Software:

  • An end-user interface for real-time data visualization and emotion annotation.

For detailed methodology and technical specifications, refer to the scientific paper available here.

Results

The EmotiphAI platform can:

• Collect data from up to 30 devices at 50Hz (1 channel), or 10 devices at 100Hz (2 channels).
• The platform was successfully used to collect a real-world dataset, comprising over 350 hours of data. This dataset is publicly available here.
• Scientific paper available here.

DEMOs

Installation

Installation can be easily done with the Clone or Download button above:

bash $ git clone https://github.com/PatriciaBota/EmotiphAI.git

Configuration

• Configurations can be found at fastapi/src/core/config.py

Run

1. make create-venv
2. make install
3. make run

To get started with EmotiphAI:

1. Set up the local infrastructure with the required hardware and software.
2. Deploy the wearable devices to participants.
3. Use the platform's interface to monitor and annotate data in real-time or retrospectively.

Acknowledge

This work was funded by FCT - Fundação para a Ciência e a Tecnologia under grants 2020.06675.BD and FCT (PCIF/SSO/0163/2019 SafeFire), FCT/MCTES national funds, co-funded EU (UIDB/50008/2020 NICE-HOME), Xinhua Net FMCI (S-0003-LX-18), Ministry of Economy and Competitiveness of the Spanish Government co-founded by ERDF (TIN2017-85409-P PhysComp), and IT - Instituto de Telecomunicacações, by the European Regional Development Fund (FEDER) through the Operational Competitiveness and Internationalization Programme (COMPETE 2020), and by National Funds (OE) through the FCT under the LISBOA-01-0247-FEDER-069918 “CardioLeather” and LISBOA-1-0247-FEDER-113480 “EpilFootSense”.