基于多任务级联卷积网络模型的人脸检测和识别

Face Detection and Recognition Based on a Multitask Cascaded Convolution Network Model

深度学习的集成特征提取这一优点使得它广泛应用于人脸检测和识别。提出了一种多任务级联卷积网络模型(Multitask Cascaded Convolution Network,MTCNN)。基于Tensor Flow平台,基于改进的任务级联卷积网络模型检测到人脸,并且用Face Net算法对人脸进行特征提取,用KNN算法对人脸进行识别。实验结果表明,对不同光照下多人图像和遮挡图像的人脸进行检测和识别,具有良好的鲁棒性。

The integrated feature extraction is an advantage of deep learning, which makes it applicable to face detection and recognition. This paper proposes a multitask cascaded convolution network model. Human faces are detected by MTCNN model in Tensor Flow platform, and Face Net algorithm is used to extract features of human face, and human face is recognized using KNN algorithm. Experimental results show that this model has good robustness in detecting and recognizing many faces of images and occlusion image.