EMOTIC

EMOTIC数据集是由加泰罗尼亚开放大学创建的，包含23,571张自然场景下的图像，标注了34,320个人的情感状态。数据集内容丰富，涵盖多种情感类别和连续情感维度，如激活度、愉悦度和支配度。创建过程中，研究者通过亚马逊的Mechanical Turk平台收集标注，确保了数据的质量和多样性。EMOTIC数据集主要用于情感识别研究，特别是在自然环境下的情感状态自动识别，为机器理解和模拟人类情感提供了重要资源。

The EMOTIC dataset was created by Universitat Oberta de Catalunya. It consists of 23,571 images captured in natural scenes, with emotional states annotated for 34,320 individuals. The dataset covers diverse emotional categories and continuous emotional dimensions including arousal, valence, and dominance. During its development, researchers collected annotations via Amazon Mechanical Turk, ensuring the dataset's quality and diversity. The EMOTIC dataset is primarily used for emotion recognition research, especially the automatic recognition of emotional states in natural environments, serving as a critical resource for machine understanding and simulation of human emotions.