Depth map prediction-based living body detection method and related equipment

Abstract

The invention discloses a depth map prediction-based living body detection method and related equipment. The method comprises the following steps: acquiring a first infrared map and a second infrared map for the same shot object; inputting the first infrared image and the second infrared image into a trained depth model to respectively obtain a first feature image and a second feature image; generating a target depth map according to the first feature map and the second feature map; inputting the target depth map into a pre-trained first classification model to generate a first classification result; inputting the first infrared image or the second infrared image into a trained second classification model to obtain a second classification result, and according to the first classification result and the second classification result, determining whether the subject is a living body. The invention provides the depth map generation method with higher precision, and the depth map is obtained based on the image shot by the infrared binocular camera, so that on one hand, the equipment cost is saved, and on the other hand, the accuracy of subsequent in-vivo detection based on the depth map and the infrared map can be improved.

Description

technical field [0001] The invention relates to the technical field of biometric identification, in particular to a method for detecting a living body based on depth map prediction and related equipment. Background technique [0002] With the development and progress of technology, there are more and more authentication devices based on face recognition. For example, access control systems based on face recognition, boot systems, etc. Its main principle is to collect images of the environment, and then use a certain face recognition algorithm or face recognition model to extract the face features of the collected images and compare them with pre-saved standard images to verify the human presence in the environment. identity. [0003] Since the current face recognition is carried out through photos, non-living bodies such as mobile phone faces, printed paper, and printed faces may use facial features close to real faces to make the face recognition algorithm or model judge ...

AI Technical Summary

Problems solved by technology

When there is a lot of background noise in the shooting environment, the matching is easily disturbed. Therefore, the final depth map cannot be used for live detection well, resulting in low detection efficiency.

Images