Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification
Deep learning using an end-to-end convolutional neural network (ConvNet) has been applied to several electroencephalography (EEG)-based brain–computer interface tasks to extract feature maps and classify the target output. However, the EEG analysis remains challenging since it requires consideration...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-10-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/21/8250 |
_version_ | 1797466514125750272 |
---|---|
author | Taweesak Emsawas Takashi Morita Tsukasa Kimura Ken-ichi Fukui Masayuki Numao |
author_facet | Taweesak Emsawas Takashi Morita Tsukasa Kimura Ken-ichi Fukui Masayuki Numao |
author_sort | Taweesak Emsawas |
collection | DOAJ |
description | Deep learning using an end-to-end convolutional neural network (ConvNet) has been applied to several electroencephalography (EEG)-based brain–computer interface tasks to extract feature maps and classify the target output. However, the EEG analysis remains challenging since it requires consideration of various architectural design components that influence the representational ability of extracted features. This study proposes an EEG-based emotion classification model called the multi-kernel temporal and spatial convolution network (MultiT-S ConvNet). The multi-scale kernel is used in the model to learn various time resolutions, and separable convolutions are applied to find related spatial patterns. In addition, we enhanced both the temporal and spatial filters with a lightweight gating mechanism. To validate the performance and classification accuracy of MultiT-S ConvNet, we conduct subject-dependent and subject-independent experiments on EEG-based emotion datasets: DEAP and SEED. Compared with existing methods, MultiT-S ConvNet outperforms with higher accuracy results and a few trainable parameters. Moreover, the proposed multi-scale module in temporal filtering enables extracting a wide range of EEG representations, covering short- to long-wavelength components. This module could be further implemented in any model of EEG-based convolution networks, and its ability potentially improves the model’s learning capacity. |
first_indexed | 2024-03-09T18:39:54Z |
format | Article |
id | doaj.art-2c95f71c147b4482af56f26783200594 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-09T18:39:54Z |
publishDate | 2022-10-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-2c95f71c147b4482af56f267832005942023-11-24T06:45:05ZengMDPI AGSensors1424-82202022-10-012221825010.3390/s22218250Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion ClassificationTaweesak Emsawas0Takashi Morita1Tsukasa Kimura2Ken-ichi Fukui3Masayuki Numao4Graduate School of Information Science and Technology, Osaka University, Osaka 565-0871, JapanThe Institute of Scientific and Industrial Research (ISIR), Osaka University, Osaka 567-0047, JapanThe Institute of Scientific and Industrial Research (ISIR), Osaka University, Osaka 567-0047, JapanThe Institute of Scientific and Industrial Research (ISIR), Osaka University, Osaka 567-0047, JapanThe Institute of Scientific and Industrial Research (ISIR), Osaka University, Osaka 567-0047, JapanDeep learning using an end-to-end convolutional neural network (ConvNet) has been applied to several electroencephalography (EEG)-based brain–computer interface tasks to extract feature maps and classify the target output. However, the EEG analysis remains challenging since it requires consideration of various architectural design components that influence the representational ability of extracted features. This study proposes an EEG-based emotion classification model called the multi-kernel temporal and spatial convolution network (MultiT-S ConvNet). The multi-scale kernel is used in the model to learn various time resolutions, and separable convolutions are applied to find related spatial patterns. In addition, we enhanced both the temporal and spatial filters with a lightweight gating mechanism. To validate the performance and classification accuracy of MultiT-S ConvNet, we conduct subject-dependent and subject-independent experiments on EEG-based emotion datasets: DEAP and SEED. Compared with existing methods, MultiT-S ConvNet outperforms with higher accuracy results and a few trainable parameters. Moreover, the proposed multi-scale module in temporal filtering enables extracting a wide range of EEG representations, covering short- to long-wavelength components. This module could be further implemented in any model of EEG-based convolution networks, and its ability potentially improves the model’s learning capacity.https://www.mdpi.com/1424-8220/22/21/8250brain–computer interface (BCI)electroencephalography (EEG)emotion classificationmachine learningconvolutional neural network (ConvNet) |
spellingShingle | Taweesak Emsawas Takashi Morita Tsukasa Kimura Ken-ichi Fukui Masayuki Numao Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification Sensors brain–computer interface (BCI) electroencephalography (EEG) emotion classification machine learning convolutional neural network (ConvNet) |
title | Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification |
title_full | Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification |
title_fullStr | Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification |
title_full_unstemmed | Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification |
title_short | Multi-Kernel Temporal and Spatial Convolution for EEG-Based Emotion Classification |
title_sort | multi kernel temporal and spatial convolution for eeg based emotion classification |
topic | brain–computer interface (BCI) electroencephalography (EEG) emotion classification machine learning convolutional neural network (ConvNet) |
url | https://www.mdpi.com/1424-8220/22/21/8250 |
work_keys_str_mv | AT taweesakemsawas multikerneltemporalandspatialconvolutionforeegbasedemotionclassification AT takashimorita multikerneltemporalandspatialconvolutionforeegbasedemotionclassification AT tsukasakimura multikerneltemporalandspatialconvolutionforeegbasedemotionclassification AT kenichifukui multikerneltemporalandspatialconvolutionforeegbasedemotionclassification AT masayukinumao multikerneltemporalandspatialconvolutionforeegbasedemotionclassification |