ExpW

该数据集包含了从谷歌图片下载的91,793张人脸图像，它代表了大规模的自然场景，涵盖了多样的种族群体和面部姿态。此外，该数据集被视为展示跨领域适应性的目标数据集。规模上，共有91,793张图像，任务是对人脸表情进行识别（FER）。

This dataset contains 91,793 facial images downloaded from Google Images, which represents large-scale natural scenes and covers diverse racial groups and facial poses. In addition, this dataset is regarded as a target dataset for demonstrating cross-domain adaptation. In terms of scale, it has a total of 91,793 images, and the task is facial expression recognition (FER).