CAS-PEAL是中国著名的人脸图像数据库，包含大量多视角、多种表情和光照条件下的面部图像，广泛应用于人脸识别与分析研究。
The CAS-PEAL face database was developed under the sponsorship of the National Hi-Tech Program and ISVISION by the Face Recognition Group at JDL, ICT, CAS. The objectives in creating this database include providing a large-scale Chinese face dataset for researchers worldwide to train and evaluate their algorithms; facilitating advancements in facial recognition technology through diverse image variations such as Pose, Expression, Accessories, and Lighting (PEAL); and promoting state-of-the-art face recognition technologies with practical applications specifically tailored for the Asian population. The CAS-PEAL database currently includes 99,594 images of 1040 individuals (595 males and 445 females), featuring a range of PEAL variations.
For each individual in the dataset, nine cameras are arranged evenly on a horizontal semicircle to capture different facial poses simultaneously. Additionally, subjects were asked to look up and down for another eighteen images to further diversify pose variation. The database also includes five types of expressions, six kinds of accessories (three glasses styles and three cap designs), and fifteen lighting conditions.
Currently, the CAS-PEAL face database is partially available in a subset named CAS-PEAL-R1, which contains 30,900 images from all 1040 subjects. This release is intended for research purposes only on an individual basis. JDL acts as the technical agent responsible for distributing this dataset and retains copyright over all images within it.
5星
