Method for carrying out in-vivo detection based on human face recognition

Abstract

The invention relates to a method for carrying out in-vivo detection based on human face recognition. The method comprises the following steps that a video including human faces is input, and the video is cut into picture sequences according to frame frequency; the human face in each frame of picture is detected out through the human face recognition technology; the pixel value of a green channel in a human face frame in one frame of picture is extracted to serve as the pixel value of the picture, the pixel values of all the pictures are ranked according to the sequence, and a pixel value sequence is formed; the pixel value sequence is subjected to noise reduction, then Fourier transform is carried out, and frequency domain information of the pixel value sequence is obtained; a heart rate value is calculated according to the frequency domain information; whether the human faces in the video are real human faces or not is judged according to the heart rate value. Heartbeat information reflected by the human faces is utilized, human face in-vivo detection is carried out faster in a better interaction mode, a tester does not need to carry out operation according to a voice instruction, and the detection speed is high. Hardware equipment such as infrared cameras or 3D cameras is not needed, the cost is low, and the method can be widely applied.

Description

technical field [0001] The invention belongs to the technical field of face recognition, and in particular relates to a method for live body detection based on face recognition. Background technique [0002] Face recognition is a biometric technology for identification based on human facial feature information. Liveness detection is a method for distinguishing between photos and real people, videos and real people in the face recognition system. [0003] Today's face detection mostly adopts command-based interactive methods, through voice prompts for action commands, such as blinking, shaking the head, opening the mouth, etc., to judge whether the person participating in the detection activity is a real person, and there are problems such as slow speed and uncooperative participants. The disadvantage of multispectral face detection is that it is easily affected by external light. The disadvantages of using infrared cameras or using three-dimensional image technology are hi...

AI Technical Summary

Problems solved by technology

The disadvantage of multispectral face detection is that it is easily affected by external light.

The disadvantages of using infrared cameras or using 3D image technology are high cost, demanding equipment hardware conditions, and not having wide application significance

[0004] At the same time, the current algorithm for heartbeat detection using face recognition has the problem that it can also detect heartbeats in photos or videos containing faces.

Images