Face representation attack detection method based on LBP-VAE anomaly detection model

Abstract

The invention discloses a face representation attack detection method based on an LBP-VAE anomaly detection model. The method comprises the steps of constructing a LBP-VAE anomaly detection model, obtaining a training sample, wherein the training sample only needs a real sample, extracting LBP features from the training sample to obtain a sample feature vector, taking the feature vector of the training sample as the input of VAE, and training a VAE network to obtain a complete LBP-VAE anomaly detection model, wherein when a human face representation attack sample is input into the model, the output error of the VAE network is very large and will be detected as anomaly due to the fact that sample feature space distribution is different from that of a real sample, and when the real sample isinput into the model, the output error of the VAE network is small, so that two types of samples can be distinguished. The attack detection method disclosed by the invention has good detection performances for different types of face representation attack samples, is strong in noise robustness, and can adapt to different real scenes.

Description

technical field [0001] The invention relates to the technical fields of image processing and biological security, in particular to a face representation attack detection method based on an LBP-VAE anomaly detection model. Background technique [0002] Nowadays, face recognition technology has been applied in all aspects of daily life, such as face attendance system, mobile phone face unlocking, face payment and so on. When the face becomes the key biometric feature in many identification and authentication systems, once a malicious person pretends to be a legitimate user and successfully passes the face recognition system, it will bring unpredictable security accidents and economic losses. Attempts to use legitimate user's face photos, videos and other means to borrow the user's identity through the operation of the face recognition system are called face representation attacks. The method of detecting this type of attack is called face detection. [0003] In human face li...

AI Technical Summary

Problems solved by technology

However, due to the small number of pictures in the current live detection data set and the large differences between different types of attack images, the most important thing is that there are many unknown attack types in the actual scene. here comes a big challenge

The current method mainly regards liveness detection as a binary classification task. However, from the above difficulties, the binary classifier learned from the existing small number of attack samples with large intra-class differences will not work in the face of unknown attack types. Existing experiments have shown that the performance of the binary classification method drops sharply in the face of cross-dataset testing, and it is difficult to apply it to actual scenarios

Images