WEBEmo
收藏资源简介:
WEBEmo数据集是由加州大学河滨分校电气与计算机工程系的研究团队创建,包含约268,000张高质量的库存照片,覆盖25个细粒度的情感类别。该数据集通过使用一个库存网站检索网络图像,并利用这些图像进行无额外人工标注的学习。创建过程中,研究团队利用了心理学中的情感层次模型来指导学习,确保数据集的多样性和广泛性。WEBEmo数据集主要应用于视觉情感分析领域,旨在解决现有情感识别数据集中的偏差问题,提高模型的泛化能力。
The WEBEmo dataset was developed by a research team from the Department of Electrical and Computer Engineering at the University of California, Riverside. It contains approximately 268,000 high-quality stock photos spanning 25 fine-grained emotion categories. This dataset retrieves web images from a stock photo website and leverages these images for learning without additional manual annotations. During its development, the research team adopted the hierarchical emotion model in psychology to guide the learning process, ensuring the diversity and broad coverage of the dataset. The WEBEmo dataset is mainly applied in the field of visual emotion analysis, aiming to address the bias issues in existing emotion recognition datasets and improve the generalization ability of models.




