CMOSE
收藏资源简介:
CMOSE数据集是由香港科技大学创建的一个综合多模态在线学生参与度数据集,包含高质量的标签。该数据集收集自在线演示培训课程的视频片段,涵盖了多种自然场景下的参与者行为。数据集中的每个视频片段都与由心理学专家指导的标签员分配的参与度标签相关联。CMOSE数据集不仅提供了丰富的视觉特征,如预训练视频特征、高级面部特征和音频特征,还特别关注了数据的不平衡问题和类内变异,通过MocoRank机制进行处理。该数据集的应用领域主要集中在在线学习环境中,旨在通过自动检测学生的参与度来提高学习效果。
The CMOSE dataset is a comprehensive multimodal online student engagement dataset developed by The Hong Kong University of Science and Technology, featuring high-quality labels. Collected from video clips of online demo training courses, this dataset covers participant behaviors in various natural scenarios. Each video clip in the dataset is associated with engagement labels assigned by annotators guided by psychology experts. The CMOSE dataset not only provides rich visual features including pre-trained video features, advanced facial features and audio features, but also pays special attention to data imbalance and intra-class variation, which are handled via the MocoRank mechanism. Its application scenarios mainly focus on online learning environments, aiming to improve learning effectiveness through automatic detection of student engagement.

- 1CMOSE: Comprehensive Multi-Modality Online Student Engagement Dataset with High-Quality Labels香港科技大学 · 2023年



