Non-living body attack discrimination method and device suitable for image, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a non-living body attack discrimination method and device suitable for an image, equipment and a medium. The method comprises the following steps: carrying out local nonlinear normalized image calculation, sub-image division and asymmetric generalized Gaussian distribution fitting and parameter estimation calculation on an initial face image to obtain a first parameter estimation value set; performing local nonlinear normalized image calculation, image division, asymmetric generalized Gaussian distribution fitting and parameter estimation calculation on the down-sampled face image to obtain a second parameter estimation value set; and inputting the first parameter estimation value set and the second parameter estimation value set into a target classification prediction model to carry out moire prediction and reflection prediction, and obtaining a non-living body attack discrimination result according to a classification prediction result. Local statistical characteristics are utilized to find whether moire and / or reflection of non-living body attacks exist or not, and a binocular camera is prevented from being used. The method is suitable for intelligent government affairs, digital medical treatment, science and technology finance and the like.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, in particular to a non-living attack discrimination method, device, equipment and medium applicable to images. Background technique [0002] With the development of image-based face recognition technology, non-living attack has become a common identity fraud, and the accurate identification of non-living attack has become an important factor for the wide application of face recognition technology. The existing technology uses a binocular camera to avoid the impact of non-living attacks on the accuracy of face recognition results, that is, two calibrated cameras are used to shoot, and then the depth of the face is judged according to the shooting results. Accurate identification has achieved good results. However, the cost of binocular cameras is relatively high. For some application scenarios where binocular cameras cannot be installed, for example, using mobile electro...

AI Technical Summary

Problems solved by technology

[0003] The main purpose of this application is to provide a non-living attack discrimination method, device, equipment, and medium suitable for images, aiming to solve the problem of using binocular cameras for face depth discrimination when performing non-living attack discrimination in the prior art. There is a technical problem that the cost is high and it is not suitable for application scenarios where binocular cameras cannot be installed

Images