LiveBeauty
收藏资源简介:
LiveBeauty是由阿里巴巴集团和上海交通大学联合创建的首个大规模直播场景下的面部吸引力预测数据集。该数据集包含10,000张从直播平台直接采集的人脸图像,并通过精心设计的主观实验获得了200,000条吸引力评分,使其成为直播场景下最大的开放访问FAP数据集。数据集的创建过程包括五步自动采样流程、主观实验和数据清洗方法,确保了数据的高质量。该数据集旨在解决直播场景中面部吸引力预测的挑战,应用于面部美化、内容推荐等领域。
LiveBeauty is the first large-scale dataset for facial attractiveness prediction (FAP) in live streaming scenarios, jointly created by Alibaba Group and Shanghai Jiao Tong University. This dataset contains 10,000 facial images directly collected from live streaming platforms, and has obtained 200,000 attractiveness ratings through well-designed subjective experiments, making it the largest open-access FAP dataset in live streaming scenarios. The creation of the dataset includes a five-step automatic sampling pipeline, subjective experiments and data cleaning methods, which ensure the high quality of the data. This dataset aims to address the challenges of facial attractiveness prediction in live streaming scenarios, and can be applied to fields such as facial beautification and content recommendation.





