DAiSEE

该数据集名为DAiSEE，包含了9,068个视频，这些视频是从112位在线课程中的授课者拍摄的，并对参与度以及三种其他情感状态进行了标注。值得注意的是，该数据集在参与度级别上存在高度不平衡，且包含有序分类。总视频数量为9,068个，任务是对多类别参与度级别进行分类。

The dataset is named DAiSEE, which contains 9,068 videos captured from 112 instructors in online courses, with annotations for engagement levels and three other emotional states. Notably, this dataset has highly imbalanced engagement level categories and involves ordinal classification tasks. The total number of videos is 9,068, and the task is multi-class classification of engagement levels.