SCUT-FBP5500
收藏资源简介:
SCUT-FBP5500是由华南理工大学创建的一个大型多范式面部美预测基准数据集,包含5500张正面人脸图像,涵盖男性/女性、亚洲/高加索人种及不同年龄段。数据集通过60名志愿者进行评分,评分范围为1至5分,同时包含面部特征点和美分数分布。该数据集支持基于外观或形状的面部美分类、回归或排名模型,适用于跨文化面部美分析、个性化面部美预测及自动面部美化等研究领域,旨在解决面部美预测的多范式计算问题。
SCUT-FBP5500 is a large-scale multi-paradigm facial beauty prediction benchmark dataset developed by South China University of Technology. It contains 5500 frontal facial images, covering male and female subjects, Asian and Caucasian ethnic groups, as well as diverse age groups. The dataset was scored by 60 volunteers, with the scoring scale ranging from 1 to 5, and it also includes facial landmarks and beauty score distributions. This dataset supports appearance-based or shape-based facial beauty classification, regression and ranking models, and is applicable to research fields such as cross-cultural facial beauty analysis, personalized facial beauty prediction and automatic facial beautification. It aims to address the multi-paradigm computational challenges in facial beauty prediction.




