سامانه همایش‌ها TSTA | ثبت‌نام و ارسال مقاله کنفرانس‌ها و کنگره‌های ملی

Home / سی و دومین کنفرانس ملی و دهمین کنفرانس بین المللی مهندسی زیست پزشکی ایران

Binary Discrete Emotion Detection with Peripheral and Fp1-Fp2 EEG Signals on PEEFS Dataset

Authors :

Fatemeh Shalchizadeh¹ Sina Shamekhi² Mahdi Jafari Asl³

1- صنعتی تبریز (سهند) 2- صنعتی تبریز (سهند) 3- صنعتی تبریز (سهند)

Keywords :

Emotion Detection،PEEFS Dataset،Peripheral Physiological Signals،Binary Classification،minimal EEG channels،Machine Learning

Abstract :

This study presents a binary discrete emotion detection approach that integrates peripheral physiological signals with EEG recordings from the Fp1 and Fp2 channels, using data from the PEEFS database. Features from both modalities were extracted to classify emotions in two scenarios: 1) individual emotions versus a neutral state, and 2) between discrete emotion pairs. Results indicate that peripheral signals generally outperform EEG alone, particularly in distinguishing emotions such as happiness and sadness from neutral, with F1-scores above 90%. However, EEG proves critical in specific cases; for instance, distinguishing between fear and neutral emotions achieves 88.9% accuracy using EEG alone, and reaches 100% accuracy when combined with peripheral signals. Similarly, emotion pairs involving fear, anger, and disgust benefit substantially from EEG integration, achieving perfect classification accuracy in several instances (e.g., fear-anger, disgust-anger). Overall, the combined modality consistently enhances performance across tasks. These findings highlight the complementary role of minimal EEG alongside peripheral signals in emotion recognition, supporting the development of practical and reliable systems for emotion monitoring applications.