Living body detection method and device

Abstract

The embodiment of the invention provides a living body detection method and device, and the method comprises the steps: extracting noise from a target face image based on a first convolutional neuralnetwork, and determining a first classification probability of the target face image based on a second convolutional neural network according to the noise and the target face image; Based on a third convolutional neural network, extracting an input surface depth feature of the target face image, and determining a second classification probability of the target image according to the input surfacedepth feature; And determining the authenticity of the target face image according to the first classification probability and the second classification probability. According to the embodiment of theinvention, the deception behavior of face recognition by using pictures and videos can be automatically recognized, the speed is high, the accuracy is high, the robustness is high, and the equipmentrequirement is low.

Description

technical field [0001] Embodiments of the present invention belong to the technical field of image classification, and more specifically, relate to a living body detection method and device. Background technique [0002] When using a face recognition system for face recognition, it is likely that a photo or video containing a human face will be placed in front of the face recognition system for recognition instead of real face recognition. Therefore, liveness detection is required, that is, to distinguish photos from real images, or videos from real images in the face recognition system, to prevent face spoofing. [0003] Most of today's liveness detection methods regard face anti-spoofing as a black-box binary classification problem. The method opens up the black box by modeling the process of generating spoofed images from raw real-time images. In the case of anti-spoofing interference, the spoofed image can be seen as a re-rendering of the real image with some special n...

AI Technical Summary

Problems solved by technology

Among them, the method based on interactive actions is slow in detection speed, difficult for users to cooperate, and poor in interactivity; based on 3D image modeling technology, it requires a lar...

Images