文献类型：期刊文献

英文题名：Resource-efficient cross-subject emotion recognition from electroencephalogram via spiking domain discriminators

作者：Li, Dongdong[1];Huang, Shengyao[1];Shen, Yujun[1];Wang, Zhe[1]

机构：[1]East China Univ Sci & Technol, Shanghai 200237, Peoples R China

年份：2025

卷号：162

外文期刊名：ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE

收录：;EI(收录号：20254119281455);WOS:【SCI-EXPANDED(收录号:WOS:001589250000018)】；

基金：This work is supported by National Natural Science Foundation of China under Grant No. 62276098 and Wenzhou Science and Technology Bureau under Grant No. ZS2024001.

语种：英文

外文关键词：Electroencephalogram; Emotion recognition; Spiking neural network; Domain adaptation

摘要：Real-time emotion recognition poses a significant challenge in electroencephalogram (EEG) based emotion recognition, as it requires the immediate processing of EEG data. This necessity imposes substantial demands on the model's resource consumption. To address this issue, this paper introduces a novel approach to EEG emotion recognition using a Cross Domain Spiking Convolutional Network (CDSCN), focusing on developments in the design of the spiking convolutional block. To address individual differences, the CDSCN incorporates Z-Score normalization at the feature level and introduces a spiking domain discriminator at the model level. These innovations aim to mitigate variations in data distribution across individuals and domains, thereby enhancing the model's robustness and generalizability. Additionally, the CDSCN introduces a novel pooling fusion layer within the spiking convolutional block to optimize computational efficiency while preserving discriminative performance. Experimental evaluations on two publicly available datasets validate the effectiveness of the proposed CDSCN in achieving both accurate and generalized emotion recognition.