Living body detection method and system based on double-decoupling generation and semi-supervised learning

Abstract

The invention discloses a living body detection method and system based on double-decoupling generation and semi-supervised learning, and the method comprises the steps: carrying out the data preprocessing, obtaining an original sample of an RGB color channel graph, and carrying out the pairing of the same identity, and obtaining a true and false image pair; a real person encoder outputs a real person identity vector, a prosthesis encoder outputs a prosthesis identity vector and a prosthesis mode vector, the three vectors are combined and sent to a decoder to obtain a reconstructed real and false image pair, a double-decoupling generation loss function is constructed, and noise is sent to the trained decoder to obtain a generation sample; a labeled sample, an unlabeled sample and an enhanced unlabeled sample are constructed for the original sample and the generated sample, the labeled sample, the unlabeled sample and the enhanced unlabeled sample are sent to a teacher learning module to obtain teacher semi-supervised loss, false labels of the unlabeled sample and no label loss after enhancement, and a detector and teacher network parameters are updated; determining a threshold value by using the verification set; and loading the test data to a detector to obtain a classification score, and judging a classification result according to a threshold value. According to the invention, the robustness of the living body detection model can be improved.

Description

technical field [0001] The invention relates to the technical field of face recognition and anti-spoofing detection, in particular to a living body detection method and system based on double decoupling generation and semi-supervised learning. Background technique [0002] Today, there is a dramatic increase in the use of facial biometrics in businesses and industries. For example, facial unlocking technology can be used to protect personal privacy in electronic devices, and facial biometrics can be used to authenticate payments. However, using the face as a biometric for authentication is not secure. Facial biometric systems can be vulnerable to spoofing attacks. Generally speaking, face spoofing attacks can be divided into four categories: 1) photo attack: the attacker uses the printed or displayed photo to deceive the authentication system; 2) video replay attack: the attacker uses the pre-photographed image of the victim. Video deception authentication system; 3) face ...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the defects and deficiencies in the prior art, the present invention provides a living body detection method based on double decoupling generation and semi-supervised learning, which uses decoupling learning to model living body / prosthesis features, and synthesizes generated samples in the latent space to expand The data set improves the diversity of training data. At the same time, it ensures the high discrimination of generated samples and reduces the impact of generated noise on model learning. Specifically, a technical solution based on double decoupling generation and semi-supervised learning is adopted to solve the problem of live detection. The technical problems of insufficient model data diversity and poor generalization have achieved the technical effect of effectively improving the generalization performance while ensuring the accuracy rate in the library

Images