IMEMNET

该数据集名为IMEMNet，它融合了多种图像数据集和一个音乐数据集，这些数据集都带有独特的情感倾向-唤醒度标签。具体来说，它包含了25,620张图像和23,944段音乐，进而形成了144,435组图像-音乐对。此外，该数据集还解决了多模态学习中数据稀缺的问题，包含了来自IAPS、NAPS以及EMOTIC数据集的图像，规模之大可见一斑。该数据集的主要任务是专注于情感分析的多模态研究。

The dataset is named IMEMNet. It integrates multiple image datasets and one music dataset, all annotated with unique valence-arousal labels. Specifically, it contains 25,620 images and 23,944 music segments, forming a total of 144,435 image-music pairs. Moreover, this dataset addresses the issue of data scarcity in multimodal learning, and incorporates images from the IAPS, NAPS and EMOTIC datasets, showcasing its substantial scale. The core task of this dataset centers on multimodal research for sentiment analysis.