FIRMED: A Peak-Centered Multimodal Dataset with Fine-Grained Annotation for Emotion Recognition

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Tang, Hao, Xie, Songyun, Xie, Xinzhou, Liao, Can, Li, Bohan, Tian, Zhongyu, Zheng, Dalu
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910088889892864
author Tang, Hao
Xie, Songyun
Xie, Xinzhou
Liao, Can
Li, Bohan
Tian, Zhongyu
Zheng, Dalu
author_facet Tang, Hao
Xie, Songyun
Xie, Xinzhou
Liao, Can
Li, Bohan
Tian, Zhongyu
Zheng, Dalu
contents Traditional video-induced physiological datasets usually rely on whole-trial labels, which introduce temporal label noise in dynamic emotion recognition. We present FIRMED, a peak-centered multimodal dataset based on an immediate-recall annotation paradigm, with synchronized EEG, ECG, GSR, PPG, and facial recordings from 35 participants. FIRMED provides event-centered timestamps, emotion labels, and intensity annotations, and its annotation quality is supported by subjective and physiological validation. Benchmark experiments show that FIRMED consistently outperforms whole-trial labeling, yielding an average gain of 3.8 percentage points across eight EEG-based classifiers, with further improvements under multimodal fusion. FIRMED provides a practical benchmark for temporally localized supervision in multimodal affective computing.
format Preprint
id arxiv_https___arxiv_org_abs_2507_02350
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FIRMED: A Peak-Centered Multimodal Dataset with Fine-Grained Annotation for Emotion Recognition
Tang, Hao
Xie, Songyun
Xie, Xinzhou
Liao, Can
Li, Bohan
Tian, Zhongyu
Zheng, Dalu
Human-Computer Interaction
Traditional video-induced physiological datasets usually rely on whole-trial labels, which introduce temporal label noise in dynamic emotion recognition. We present FIRMED, a peak-centered multimodal dataset based on an immediate-recall annotation paradigm, with synchronized EEG, ECG, GSR, PPG, and facial recordings from 35 participants. FIRMED provides event-centered timestamps, emotion labels, and intensity annotations, and its annotation quality is supported by subjective and physiological validation. Benchmark experiments show that FIRMED consistently outperforms whole-trial labeling, yielding an average gain of 3.8 percentage points across eight EEG-based classifiers, with further improvements under multimodal fusion. FIRMED provides a practical benchmark for temporally localized supervision in multimodal affective computing.
title FIRMED: A Peak-Centered Multimodal Dataset with Fine-Grained Annotation for Emotion Recognition
topic Human-Computer Interaction
url https://arxiv.org/abs/2507.02350