FIRMED
收藏资源简介:
FIRMED数据集由西北工业大学的研究团队创建,旨在通过即时回忆范式进行细粒度情绪标注,解决传统情绪生理数据集中粗粒度标注带来的标签噪声问题。该数据集收集了参与者在观看刺激后的即时回忆阶段所标记的情感事件时间戳、情感标签和强度,并通过生理证据和识别性能进行了验证。数据集收集了多种生理信号,包括脑电图(EEG)、心电图(ECG)、皮肤电反应(GSR)和光电容积描记图(PPG)数据,以及参与者观看刺激时的面部表情数据,提供了行为数据。数据集利用来自社交媒体平台的真实视频作为刺激,提高了生态效度。FIRMED数据集可用于情绪识别研究,以解决标签噪声问题并提高情绪识别性能。
The FIRMED dataset was developed by a research team from Northwestern Polytechnical University. It aims to perform fine-grained emotion annotation via the immediate recall paradigm, so as to resolve the label noise problem caused by coarse-grained annotations in traditional emotional physiological datasets. This dataset collects affective event timestamps, emotion labels and their intensities annotated by participants during the immediate recall stage after they viewed the stimuli, and has been validated using physiological evidence and recognition performance metrics. It gathers various physiological signals, including electroencephalogram (EEG), electrocardiogram (ECG), galvanic skin response (GSR) and photoplethysmography (PPG) data, as well as facial expression data of participants while they watched the stimuli, and also provides behavioral data. The dataset employs real videos sourced from social media platforms as experimental stimuli, which enhances its ecological validity. The FIRMED dataset can be applied to emotion recognition research to mitigate label noise issues and boost emotion recognition performance.
FIRMED数据集概述
基本信息
- 数据集名称:FIRMED
- 托管平台:GitHub
- 托管地址:https://github.com/NIlab666/FIRMED
数据集描述
(注:根据提供的README内容,该数据集未包含具体描述信息)

- 1From Coarse to Fine-Grained Emotion Annotation: An Immediate Recall Paradigm with Validation through Physiological Evidence and Recognition Performance西北工业大学 · 2025年



