EmoGator
收藏资源简介:
EmoGator数据集由佛罗里达大学的Fred W. Buhl创建,包含32,130个来自357位发言者的非言语情感发声样本,总时长16.9654小时。每个样本被分类到30个不同的情感类别中。数据集通过志愿者和众包工作者收集,使用文本提示帮助引发情感反应。EmoGator旨在解决语音情感识别中的数据稀缺问题,特别是在非言语情感发声方面,为机器学习分类方法提供基准。该数据集的应用领域包括视频游戏中的非玩家角色响应、早期儿童教育、社交机器人交互等,旨在提高计算机对人类情感表达的理解和响应能力。
The EmoGator dataset was developed by Fred W. Buhl of the University of Florida. It comprises 32,130 nonverbal emotional vocal samples from 357 speakers, with a total duration of 16.9654 hours. Each sample is classified into 30 distinct emotional categories. The dataset was collected through volunteers and crowd workers, utilizing text prompts to elicit emotional reactions from participants. EmoGator is designed to address the problem of data scarcity in speech emotion recognition, particularly for nonverbal emotional vocalizations, and serves as a benchmark for machine learning-based classification approaches. Its potential application domains include non-player character responses in video games, early childhood education, social robot interaction, and other fields, with the overarching goal of improving computers' capacity to comprehend and respond to human emotional expressions.




