相关数据集
Human-interaction-based Emotion Forecasting (Hi-EF)
Hi-EF数据集由复旦大学开发，专注于基于人际交互的情感预测任务。该数据集包含3069个多层次情境交互样本，每个样本包含丰富的情感相关标签和三种模态信息。数据集的创建过程包括候选样本生成、异常检测和多方面辅助的可靠标注。Hi-EF数据集主要应用于情感计算领域，旨在通过分析交互情境中的情感动态，预测未来情感状态，为情感管理和支持提供新的工具和方法。
arXiv2024-07-23 更新850
Studying the Aesthetics of Images and Advertising Films
The notion, that aesthetic objects are not only things but also does something to those who engage with them, is consistent among various researchers in the arts and humanities; and also that this is
DataCite Commons2020-07-15 更新300
A standardised database of Chinese emotional film clips
Film clips are widely used in emotion research due to their relatively high ecological validity. Although researchers have established various film clip sets for different cultures, the few that exist
NIAID Data Ecosystem180
数据堂—1,142人会议场景表情识别数据
1,142人会议场景表情识别数据，包含黄种人、白种人、黑种人、棕色人种多任务，年龄以中青年为主，采集了多种室内办公场景，涵盖会议室、咖啡厅、图书馆、卧室等，每位采集者采集了正常、高兴、惊奇、悲伤、愤怒、厌恶、恐惧，7种表情
魔搭社区2026-07-14 更新380
