Emotion Classification from EEG Signals Using Time-Frequency-DWT Features and ANN

• 作者: Adrian Qi-Xiang Ang Wee Wee Yi Qi Yeong
• 来源期刊: 电脑和通信(英文)
• 学科: 医学
• 关键词: EEG EMOTION Classification FEATURE Extraction
• 年，卷（期）: 2017,（3）
• 页码范围: 75-79
• 页数: 5页
• 分类号: R73

摘要

This paper proposes the use of time-frequency and wavelet transform features for emotion recognition via EEG signals. The proposed experiment has been carefully designed with EEG electrodes placed at FP1 and FP2 and using images provided by the Affective Picture System (IAP), which was developed by the University of Florida. A total of two time-domain features, two frequen-cy-domain features, as well as discrete wavelet transform coefficients have been studied using Artificial Neural Network (ANN) as the classifier, and the best combination of these features has been determined. Using the data collected, the best detection accuracy achievable by the proposed schemed is about 81.8%.