datasets for mask face recognition Since deep learning models usually require large-scale datasets for training, the lack of mask face datasets will make it difficult for the model to learn the feature mapping between whether the face is occluded by the mask, resulting in poor recognition rate. For this problem, this paper use a mask face image synthesis algorithm based on key point detection to generate the corresponding dataset, and generates a more realistic mask face image for model training through face detection, key point detection and mask coverage.
Here are our datasets for training and testing:
train dataset
VGG-Face2——masked link：https://pan.baidu.com/s/1hJ5eak6rZ00oPH4S0ad6Yg Extraction code：1sjw
test dataset
lfw_masked link：https://pan.baidu.com/s/1DMW2_aGsHQPU6VoXVgxFxg Extraction code：qa7o
CASIA-FaceV5_masked link：https://pan.baidu.com/s/1AtzCTNpKWEJzLorfvdE8zA Extraction code：d8gf