Related Experiment Video
Updated: Nov 18, 2025
10:03
Conformable Wearable Electrodes: From Fabrication to Electrophysiological Assessment
Published on: July 22, 2022
4.7K
Emotion Recognition Based on Skin Potential Signals with a Portable Wireless Device
Shuhao Chen1, Ke Jiang1, Haoji Hu1
1College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China.
Sensors (Basel, Switzerland)
|February 5, 2021
Summary
This study introduces a portable device for emotion recognition using skin potential (SP) signals. The method achieves 75% accuracy, demonstrating a feasible, cost-effective approach for recognizing emotions.
Area of Science:
- Physiology
- Computer Science
- Artificial Intelligence
Background:
- Emotion recognition is crucial for AI, robotics, and medicine.
- Existing methods often require complex and expensive equipment.
- Skin potential (SP) signals, correlated with emotions, have been underutilized due to limited research.
Purpose of the Study:
- To propose and validate a novel emotion recognition method based solely on skin potential (SP) signals.
- To develop a portable, wireless device for measuring SP signals.
- To assess the feasibility of using SP signals for recognizing four basic emotions.
Main Methods:
- A portable wireless device was created to measure skin potential (SP) between the middle finger and left wrist.
- A video induction experiment stimulated happiness, sadness, anger, and fear in 26 participants.
- 29 features were extracted from 397 SP emotion samples, classified using eight algorithms.
Main Results:
- The Gradient-Boosting Decision Tree (GBDT), Logistic Regression (LR), and Random Forest (RF) algorithms achieved the highest classification accuracy of 75%.
- This accuracy is comparable or superior to methods utilizing multiple physiological signals.
- The study successfully classified four distinct emotions based on single-channel SP data.
Conclusions:
- The skin potential (SP) signal is a viable and effective biosignal for emotion recognition.
- A portable, single-signal-based approach offers a cost-effective alternative to multi-signal systems.
- This research supports the integration of SP signals into broader emotion recognition frameworks.
