基于人脸关键点的交互式活体检测OA

With the rapid development of artificial intelligence technology,face recognition technology has been widely used,especially in the financial field,which greatly saves the cost of human resources.Aiming at the problem that face recognition is easy to be cheated by illegal elements using images and models to steal user property,this paper proposes an interactive dynamic living detection method.This method first calls the dlib library to detect the human face location in the image and locate the facial 68 points points,by affine transformation deflection image correction and optimization of image quality based on pixel operator,ran-dom combination blink of an eye,mouth,nodding and shaking the head to verify whether the recognized face is a live one,effective-ly improve the security of face recognition.Compared with other methods,the experiment shows that the error rate of the proposed method is 1.483%on the CASIA FASD dataset,5.721%on the OULU-NPU dataset,and 0.734%on the rereplay-Attack dataset.The proposed method performs better than other methods in complex environment and lighting conditions.It has stronger generaliza-tion ability and better robustness.