monster-monash/DREAMERV
收藏资源简介:
DreamerA数据集是一个多模态数据集,包含在使用音频-视觉刺激引发情感时记录的脑电图(EEG)和心电图(ECG)信号。该数据集使用14通道Emotiv EPOC耳机以128 Hz的采样率捕获。它由23名参与者在每次刺激后对自己的情感状态(效价、唤醒度和支配度)的自我评估组成。本研究中的分类任务专注于觉醒和效价标签,分别称为DreamerA和DreamerV。处理后的数据集包含170,246个长度为256的时间序列,每个时间序列代表2秒钟的数据。标签经过二值化处理,低于3的值分配给类别1,3或更高的值分配给类别2,并且已根据参与者分成交叉验证折。
DreamerA dataset is a multimodal dataset that includes electroencephalogram (EEG) and electrocardiogram (ECG) signals recorded during affect elicitation using audio-visual stimuli. The dataset was captured using a 14-channel Emotiv EPOC headset at a sampling rate of 128 Hz, consisting of self-assessments of affective states (valence, arousal, and dominance) from 23 participants after each stimulus. The classification task in this study focuses on the arousal and valence labels, referred to as DreamerA and DreamerV respectively. The processed datasets both consist of 170,246 multivariate time series each of length 256, representing 2 seconds of data per time series at a sampling rate of 128 Hz. The labels have been binarized, with values below 3 assigned to class 1 and values of 3 or higher to class 2, and the dataset has been split into cross-validation folds based on participants.




