基于掩膜生成网络的遮挡人脸检测方法

Occluded Face Detection Method Based on Mask Generation Network

针对复杂遮挡条件下人脸检测精度低的问题,提出一种基于掩膜生成网络(MGN)的遮挡人脸检测方法。对人脸训练集进行预处理,将训练人脸划分为25个子区域,并为每个子区域分别添加遮挡。将一系列添加遮挡的人脸图像和原始人脸图像作为图像对依次输入MGN进行训练,以生成对应各个遮挡子区域的遮挡掩膜字典。通过组合相关字典项生成与检测人脸遮挡区域相对应的组合特征掩膜,并将该组合特征掩膜与检测人脸深层特征图相点乘,以屏蔽由局部遮挡引起的人脸特征元素损坏。在AR和MAFA数据集上进行实验,结果表明,该方法的检测精度高于MaskNet、RPSM等方法,且检测速度较快。

To address the low accuracy of occluded face detection,an occluded face detection method based on Mask Generation Network(MGN)is proposed.The method pre-processes the dataset for face training by dividing each face image into 25 sub-regions and adding occlusions for each sub-region.Then the occluded face images and original face images are sequentially input into the MGN for training to generate an occlusion mask dictionary corresponding to each occluded sub-region.Next,related dictionary entries are combined to generate a combined feature mask corresponding to the occluded region of the to-be-detected face.The combined feature mask is multiplied with the deep feature map of the to-be-detected face to shield the facial feature element damage caused by local occlusion.The experimental results on AR and MAFA datasets show that the proposed method exhibits a higher detection accuracy than MaskNet,RPSM and other methods while keeping a high detection speed.